Email Robert |
Week of 28 August 2000 Friday, 18 April 2003 07:57 A (mostly) daily journal of the trials, tribulations, and random observations of Robert Bruce Thompson, a writer of computer books. |
|
wpoison
Search [tips]
|
Jump
to most recent update Monday, 28 August 2000 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] We had a bit of a lie-in this morning. I woke up at 0730, and the dogs were still asleep. That's much better than the usual 0630 starting time for the dogs, who begin prancing around, snouting us, and licking our faces until we're awake. I remember the days when I considered 0730 early and sometimes slept until 0830 or 0900 on weekends. Now I consider 0730 a lie-in. Border Collies certainly do take charge of one's life. I spent most of yesterday cleaning up my office in preparation for selling the house. While doing that, I rearranged my desk completely and installed the Belkin OmniCube KVM switch. Some initial impressions:
The exception I mentioned is that there is some strangeness with the Windows Me box. Most times, when I switch to the WinMe box and move the mouse, a right-click menu or two pops up, as though spurious phantom mouse clicks were being passed to the box. And, although power management is disabled on the WinMe box, about every fifth or eighth time I switch to the WinMe box, the screen remains blank and nothing I do with the keyboard or mouse fixes that. I can fix the problem by switching to a different box, moving the mouse, and then switching back to the WinMe box, whereupon everything works again. Also, perhaps every tenth time I switch to the WinMe box, the mouse is completely dead and I have to restart the system to bring it back to life. And, in a truly strange turn of events, the WinMe screensaver doesn't work. That is, no matter what the screensaver settings on the WinMe box, the screensaver never kicks in. In fairness to Belkin, I should note that WinMe is not listed as a supported operating system. Nor is the IntelliMouse Explorer which I'm running listed as a supported mouse by Belkin (although they do list the IntelliMouse and IntelliMouse Pro). Also, I have not yet installed the driver software for the IntelliMouse Explorer on any of the systems involved. So it's quite possible that installing the driver software may solve most or all the problems under WinMe. But the OmniCube works flawlessly with Windows 98 SE and Windows NT4, and also worked perfectly during the limited testing I did with a Windows 2000 box connected. I can't endorse the Belkin OmniCube as yet, because I recommend products only after using them extensively. But, despite the teething pains I've experienced with WinMe, I expect the Belkin OmniCube to eventually become a recommended product for our hardware books. If you need a KVM switch right now, you could do much worse than the Belkin OmniCube. Barbara shot this photo while I was disconnecting stuff in preparation for installing the Belkin KVM switch. I'm wearing something that everyone who works on PCs should have--an L. L. Bean headlamp. When you're working with a flashlight and find yourself needing a third hand, this fits the bill. I didn't mention it in PC Hardware in a Nutshell because I thought people might take exception to me recommending a $50 flashlight
Here's what my desk looks like after installing the Belkin OmniView KVM switch. I'm down from four keyboards, four mice, and two computers on my desk to two keyboards, two mice, and zero computers. Things are considerably less cluttered than they were before.
From left to right, the computers are:
Kiwi connects directly to the stuff on the right side--an Hitachi SuperScan Elite 751 monitor, a Microsoft Natural Keyboard Pro, and a Microsoft IntelliMouse with IntelliEye. The three other computers connect via the Belkin KVM switch to the stuff on the left side--a Dell 17" monitor, the Microsoft Natural Keyboard, and the Microsoft IntelliMouse Explorer. For the first time in a long time I feel like I have room to work. -----Original Message----- So what makes a monitor with a greater than 100% failure rate (the NECs) better than the Samsungs with a close to 0% failure rate? The several week repair time by NEC wasn't anything to rave about. Samsung have NEC beat hollow on that, and to me and my clients, it's a very important criterion. The picture quality was from visual inspection identical. Maybe there was a measurable difference, but it would not have surprised me if that had gone Samsung's way. Given that the supplier sold the NECs to us for the same price as the Samsungs, I suspect that his buy price could not have been much higher. This is all based on my personal experience. I have since discussed this with my monitor repair person. He says that NEC's lower spec monitors are a pain. They do not use standard parts and NEC wouldn't know what customer service was. For reliability, Samsung is right up there with Hitachi, Sony and Mitsubishi, as is Philips. Could it be that in your American throwaway society reliability doesn't rate? Here we like to get at least 4-5 years use from a monitor, but then they cost significantly more here than over there. My Sony G400 cost me slightly more than $US900, about the same as I paid for the Philips 17b 4 years ago. You obviously ran into a bad batch of NEC monitors. That can happen to any company, but happens more seldom with better, more expensive monitors. I will say again that the monitor market is so competitive that there aren't any bargains out there. You get (at most) what you pay for. If a given monitor is significantly less expensive than another model with similar specifications, you can bet that the first monitor has had corners cut somewhere. My own experience with NEC monitors has been uniformly good, based on dozens of monitors over 15 years or more. I remember only one NEC monitor failing, and that was at least five years old, probably more. I remember a dozen other NEC monitors that lasted seven to ten years. They were eventually discarded even though they still worked properly simply because they were too small or had too narrow a sync range to be useful any longer. Conversely, my experience with Korean monitors (Samsung, Goldstar, and Hyundai) has not been good. Most of them were adequate when new, but had much higher failure rates than Japanese monitors, both initially and as time passed. And, no, I'm not a believer in the throwaway society. Far from it, in fact. That's why I recommend buying good name brand monitors like those produced by Hitachi, Mitsubishi, NEC, and Sony. They'll last a lot longer than lesser monitors. -----Original Message----- What then if he defaults, that is does not appear or send a representative to the trial? Is he guilty by default? Unfortunately that may be the only possible way to ensure appearance. If appearance is necessary to a valid conviction and there is no penalty for non-appearance then any rational person will not appaear. I am not sure that as the penalty for non-appearance is a desireable result. But back to the basic inquiry, why can the state not demand a surety to require appearance after arrest? It seems to me that the state (or society if you prefer) has an interest as a group in enforcing a system that protects persons and their property from harm. To be effective at acheiving that end, the system must be fair, and must be seen as fair; the latter perhaps more important than the former. If the system must be seen as fair, then is it not necessary that the processes of the system be see-able, that is visible to the members of the society? A public trial, the essential core process, does not happen if the accused is a no-show, or if his non-appaearnce is taken by the system as an admission of liability. Thus, it seems to me that the system must have a mechanism, other than blind trust in the accused to appear or the assumption of liability if he does not appear, for the system to perform the function (punishment and deterrence of "criminal" actions) the society requires of the system. Thus, the bail requirement - you will show for trial either because we have you b/c you have insufficient resources to be trusted to appear or your resources are at risk by a bail agreement and thus you choose to appear. RAS As an aside, experience has convinced me that one of the truly great evils of our current system is that it too often labors unseen by the memebers of the society whose property and persons it exists to protect. Hmm. I thought I'd covered that. The accused can show up for his trial or not, as he pleases. At that point, he's still presumed innocent, and the state and the court should exercise no constraints on him. But If he chooses not to show up, he's leaving the prosecution unanswered, so it's clearly in his interest to show up or send a representative to defend himself against the charges. But until he is convicted, the state has no right to demand his appearance or to require any type of surety for that appearance. In law, he is no different than you or I. Nor did I say that nonappearance should be taken as an admission of guilt. In practice, under such a system, nearly every accused would show up for his trial, simply because he couldn't afford to allow the prosecution to make its case unanswered. Worse still, the threat of being outlawed should effectively compel any accused to show up for punishment if he is convicted. Wolfsheads may, after all, be shot on sight by anyone. So you probably don't want to be a wolfshead. -----Original Message----- Luis Bazdresch wrote: "If high-quality slide scanners become as cheap as a good digital camera, I would still consider getting one of those instead of the camera." That day is already here as far as I can tell. I'm very pleased with my CanoScan FS 2710 and the 2700 dpi scans are certainly higher resolution than all but the very top end digital cameras that cost several times as much. Printing AFAIK remains problematic. The inks used by run-of-the-mill inkjet printers are expensive and not very permanent. There are lightfast inks available, but they are probably still ruinously expensive compared to conventional photographic printing. The following URL may help you decide: [here] Well, ruinously expensive compared to what? Inkjet consumables certainly aren't cheap--a point I've made repeatedly--but then neither are standard color prints. Nor are standard color prints particularly immune to fading from exposure to light. That's a problem that plagues any dye-based (versus pigment-based) process, including color negative, print, and slide materials based on color couplers. That's why I prefer processes that do not use color couplers, such as Kodachrome, Cibachrome, and dye transfer (for which, alas, Kodak discontinued the matrix film and dyes). Thanks for the URL. I'm accumulating data on photo printing now. -----Original Message----- Robert "PC Magazine and others typically do superficial side-by-side comparative reviews, because they're easy to do and that's what readers seem to want. So PC Magazine contacts a bunch of monitor makers, who send them examples of their monitors. PC Magazine then sets up the monitors, does some quick tests, and sends the monitors back. I don't regard such short-term reviews as being worth much. I was quite pleased with my Mag Innovision monitors when I bought them, for example, but within a year their display quality started to degrade significantly. That's the kind of thing that short-term reviews miss. When I recommend a product, it's because I've used it day-in-day-out for weeks or months." Also, the manufacturer can make sure the items are checked over before shipping. My daughter bought a system mail order. When it arrived the sound didn't work. I found that the sound card was not in the pci slot. I am sure that that sloppy workmanship would never have made it to PC Magazine. Mike Boyle There's always that danger, of course. And it's one that I'm subject to as well because a lot of the stuff I use is provided by the manufacturer (or, more likely, their PR firm). I don't worry too much about that, though. A vendor might think it worthwhile to tweak something they were sending to PC Magazine, but I don't think they'd bother for me. Most of the new stuff is standard retail-boxed product anyway, pulled straight from the warehouse and shipped out. And the flip side of that is that many review units are already well-used by the time reviewers receive them. For example, if I request a review unit of, say, a Kodak digital camera, it'll probably show up a week or a month later, with a note saying I can keep it for 30 or 60 days or whatever. Chances are good that at least one (and probably several) other reviewers have had the camera before I got it. In fact, sometimes the PR agency will ask if we can pack up an eval product and address it directly to the next reviewer in line rather than returning it to the vendor or the PR firm. -----Original Message----- "I'll know once the printer arrives and I have a chance to work with it. " I'd be interested to know what you find. I've never done any "in depth" testing, but own brief experience with an Epson 740, at 1440x720 on photo glossy paper left "dithering", which was highly noticeable on light areas (such as light skin tones) of the picture. That said I was not looking to benchmark, and so played very little with settings, paper etc. etc. ATB. Phil -- Time allowing, I'll certainly keep everyone posted about what I learn. Going in, my expectations are that the Epson 760 will produce 4X6 inch prints from a 1280X960 image that are effectively indistinguishable from standard color prints; that those prints will have reasonable light stability (probably about the same or perhaps a bit better than standard color prints); and that they'll cost about the same or perhaps a bit less than standard drugstore prints. But we'll see. -----Original Message----- Robert I don't think they would bother with a item like a camera or a card, but for a whole system with many parts I'm sure they would. You may well be right there. That's not an issue for me, because I don't review complete systems. I've never seen the point, because manufacturers switch components without notice. That's particularly true for high-volume direct vendors like Dell and Gateway. They sell in such large volumes that they often second- and third-source components and use them interchangeably. I remember once getting two apparently identical Gateway systems with nearly sequential serial numbers. When I opened them up, they had different motherboards, different hard disks, and different video cards! So I've never seen the point to reviewing an entire system when even if you ordered that "same" system immediately you might end up with one built with different components. -----Original Message----- Robert That must be why they insist on the serial number when you call them. Might be the same for cards. Consumer reports does it the right way. They send people out to buy the stuff just like a normal consumer woulod. Oh, sure. Gateway, Dell, and the other big direct makers have everything barcoded and scanned into their databases. When you give them the system serial number, they can call up the configuration on their computers and know exactly what is installed in the system, or was when it left the factory. As far as Consumer Reports, I don't pay much attention to them. -----Original Message----- Robert, Indeed, my pal has a large number of tracks from his own CDs. He also has quite a number of free tracks from the net (not all from legitimate sources, but a surprisingly large number are). A lot are from acts that NEVER have released a CD. He can put together a "Hey, listen to this in the car" compilation over a lunch hour, and leave plenty of time for eating, too. Having tried the "one from this CD, two from that one" compilation methods, I know I don't have the patience to do it for other than VERY SPECIAL people, VERY INFREQUENTLY. But even I can click select 15 songs from an MP3 list, click a button and be done with the work. But quality is the victim of convenience. On the other hand, I rarely listen to records (oops, showing my age) CDs in other than background mode. His stuff is perfectly fine for me. Well mostly. I don't like most of the acts I've never heard of. But that's a WHOLE other complaint [G]. -----Original Message----- I really enjoy your website....it's a nicely balanced mixture of hardware and other observations that shows up earlier in the day than Tom's or Jerry's (the latter of which sometimes doesn't seem to get refreshed/updated on his server for several days at a time (from my perspective), for example it's 1/2 hour before Monday morning begins and Chaos Manor stills says "Friday"), but I digress. The only thing I want to say about your monitor recommendations is that it is my gut feeling that it is probably getting overly simplistic to insist that you can *always* buy anything a manufacturer makes and sells (i.e. in the larger context of all the different models ...aimed at different markets/purposes) by *brand*. I remember how in photography, the loyalists who loved to have the camera bag filled with straight Nikkor or Canon lenses...ignoring the fact that there are superior lenses in some focal length/configurations made by makers like Tamron, etc. Still, there are reasons why a pro would want straight Nikkors, etc. ... the consistent feel (and familiarity of controls) of each piece in the arsenal, etc. So the metaphor does not parallel the monitor situation...unless one has a bank of monitors, perhaps ;-). I admit that I don't have the experience that you do in this field. I am quite happy with my 19" Mag DJ800 (Hitachi 0.26 tube) that is a demo unit that I got with an extended warrantee. We'll see what kind of tune I'm singing in say 3 years from now ;-) All the best, -Ron Snider- Thanks for the kind words. As far as monitor brands, I'll emphasize again that you get no more than you pay for. I'm not saying that every example of an Hitachi, Mitsubishi, NEC, or Sony monitor is good nor that every Mag or Princeton monitor is bad. Far from it. It's quite possible to get a bad example of a "good" brand-name monitor, just as it's quite possible to get a superb example of a monitor from one of the lesser brand names. But the odds strongly favor getting a good monitor when you buy a premium-price model and a mediocre monitor when you buy a "value" brand. Just as with camera lenses, part of what you're paying for is quality control. I'm sure that Nikon ships an occasional bad Nikkor, just as I'm sure that Tamron or Vivitar occasionally ships a superb example, but overall it's likely both that the average quality of the Nikkors is higher than that of the Tamrons or Vivitars and, just as important, that the standard deviation of quality is much smaller with the Nikkors than with the aftermarket brands. The way to insure both high average quality and small deviations in quality is to spend a lot of money on QC, and that translates to a higher selling price. Same thing with monitors. There are much larger quality variations among the "value" brands, simply because the manufacturer doesn't spend as much money on QC as the premium manufacturers do. All of that said, monitors are unfortunately the one PC component most subject to shipping damage. It is quite possible for a monitor to leave the factory in perfect working order and arrive at your doorstep having been banged around so much that it now has mediocre display quality and may also have reduced life expectancy. But, short of going to the factory and carrying the monitor home yourself, there's not much to be done about that.
|
|
wpoison
Search [tips]
|
Tuesday, 29 August 2000 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] Ahem. Barbara points out that when I published the photo of my now-neat desk yesterday, I didn't mention the rest of my workspace. My credenza, alas, is a less happy sight than my desk.
There are, believe it or not, eight computers more-or-less visible in this picture. The six across the bottom are obvious. From left to right, they are:
At the left edge of the folding table, two more systems are (kind of) visible. The one with the old towel on top and wires hanging out is an Intel D810EMO system with a Pentium III/733. The eighth system is that piece of plywood, which has some kind of motherboard (I can't remember which and I don't want to risk an avalanche to find out) and a Pentium III/866 processor (I think). I used to refer to this table and the antique inlaid wood table that sat between it and my main desk as the "Augean Tables". But now Barbara has reclaimed the antique table, with some justifiable tut-tutting I might add, so I now have only the Augean Table, singular, which doesn't sound quite right. I thought perhaps the problems with the WinMe box on the Belkin KVM switch might be due to the fact that I hadn't installed drivers. So I went over to Microsoft to download the latest mouse drivers (4 MB) and keyboard drivers (18.5 MB!). I installed them, and for a time it seemed that doing that had in fact solved the problem. But that was just coincidence. The problems showed up again, same as ever. I think the problems may be caused by WinMe itself, or perhaps by something odd about the PC hardware. So I'm going to blow away WinMe on that box and install Windows 2000 Professional. We'll see what happens. Still more review items are stacking up. Two products are at the top of the pile: Turtle Beach Santa Cruz sound card. I've used Turtle Beach sound cards for years, and have always found them preferable to Creative Labs sound cards. In particular, the recent Aureal-based Turtle Beach products with their A3D 2.0 3D positional audio support far outperformed Creative's less ambitious EAX technology. But making sound cards is a tough business to be in right now, with so many motherboards now having embedded audio. That was doubtless a major reason behind the financial problems Aureal experienced earlier this year, and it appears that Voyetra/Turtle Beach is hedging its bets by producing the new Santa Cruz, which is based on a Cirrus Logic chipset rather than one from Aureal. With very impressive specifications (although A3D 2.0 support is no longer present) and an SRP of $100, the Santa Cruz looks like it may provide very serious competition for the higher-end Creative Labs products. We'll see. It goes in the SE440BX-2V test-bed system the next time I work on it, which will probably be next weekend.StorCase Technology Data Express DE100 Removable Subsystem. StorCase Technology is a new company, spun off in June from Kingston Technology of RAM fame. The Data Express units are a combination of drive carrier and receiving chassis that allow standard 3.5" hard drives to be easily added to or removed from a system. Because its construction permits front access, the Data Express can also be used to swap front-loading 3.5" devices such as tape drives. The Data Express is available in many variants, including ones that support ATA/33, ATA/66, and various SCSI flavors. I have one of the DE100 models, which supports ATA/66, along with a couple of spare carriers. The target market for these devices is probably people who maintain and transfer huge data sets and therefore frequently need to swap hard drives in and out of their systems. In my case, though, I intend to use the DE100 as an alternative to multiple-booting the SE440BX-2V test-bed system. I've never been particularly happy with multi-boot systems. In my case, I use multi-boot to minimize the number of machines I have on and around my desk rather than to avoid having to build or buy another system. Using the DE100 instead will eliminate the problems that inevitably crop up in a multi-boot test-bed environment. I can allocate one hard drive and carrier each to Windows 98 (or WinMe), Windows NT4, and Windows 2000. Only one of those hard disks will be installed at any one time, which to all intents and purposes converts a single test-bed system into three dedicated systems. Here's an interesting Cringely (InfoWorld) column that talks about Microsoft's own use of Windows 2000 versus various Unix versions. If this stuff is true, and I suspect it is, I've been right all along in my criticisms of Windows 2000. There's more interesting stuff in some of the InfoWorld columns this week, including an Ed Foster column about the tax hazards of co-locating a web server in another state and a Nicholas Petreley column which mentions the underwhelming rate at which Windows 2000 is being adopted. Petreley says, in part:
Now, granted, Petreley is a Linux advocate and no fan of Microsoft or Windows, but I suspect he's probably more right than wrong here. -----Original Message----- Robert; On the question: "Why sacrifice audio quality by converting the MP3s back to CD-DA?" I download the Byte Audio Review and Geeks In Space to burn to CD for listening to in the car when driving. I've never managed to just sit at a PC and just listen to a audio feed like that where drive time is usually a good time to listen to them. Ah, good point. That's an application I hadn't thought about. -----Original Message----- Robert, Just saw your latest on the Belkin KVM switches and would like to contribute the following. At work we have Belkin KVM switches and they have been driving us crazy in losing the mouse when being switched, and also not being able to start machines unless the switch is switched to that particular box. The second problem is no big deal to us in the office because of the rarity that the machines get restarted, but the first is unforgivable. Learning from this experience I bought a Connect-Tek (www.connect-tek.com) switch for my home office. It works perfectly except it does not recognize the two additional buttons on the Intellimouse Explorer. It cost slightly more than the Belkins but came with excellent cables included, and also has sufficient room on the back to connect them all (unlike the Belkin). Interesting. My own experience so far is that the Belkin switch works flawlessly with Windows NT 4 and Windows 98 SE, but has all kinds of problems with Windows Me. I'm going to blow away the WinMe installation and put Windows 2000 Professional on that system to see how it does. I suppose it's possible that it's something about the hardware itself that the Belkin doesn't like. If so, running W2KP on it should turn that up. -----Original Message----- The simplest way to complie audio CD's is with the audio cd mp3 studio program. It is available as advertising supported shareware, or you can register it and dump the adware part. In any event, it will create MP3's from audio cd's, burn audio cd's from multiple exisitng audio cd's, from mp3's, from wav's etc. simple, and any idiot can make disks with it. [here] Thanks. That's one I'm not familiar with. And I confess that I'm not likely to try it, because I don't use programs that use adware and I prefer not to pay up front for a program like this before I've seen what it can do. -----Original Message----- hey Robert....i'm one of the guys who bought the cc820 motherboard and as i read your article and others decided to go back to the shop that sold me the computer......my problems started with a certain hissy noise that is produced every time i click on Start button and the menus come popping out.....with every menu that pops out, there was this unusual sound that's like a wizzzzz( if one has desktop themes enabled)!!! then my major problems were during games which simply halted....hanged!!!!!! 3 of the games that i played all caused a system hang including nfs, nba 2000..... now my question is that i asked them to install the intel 440bx-2...... i had 64 MB of SDRAM. do these need to be changed?? can they be installed on the 440bx-2? are they corrupted and does the bx-2 require RDRAM or SDRAM?? Hmmm. I'd have probably taken Intel up on their offer to swap the CC820 for a VC820 motherboard with 128 MB of RDRAM, but the SE440BX-2 is certainly a competent motherboard. To answer your questions, the SE440BX-2 uses SDRAM rather than RDRAM, the SDRAM you had on the CC820 should work fine on the SE440BX-2, and (unless the memory is mishandled) there's no reason why the CC820 would have done any damage to it. -----Original Message----- A couple of comments re: Belkin's OmniCube KVM switches - -I've had similar boot problems - occasionally if the switch isn't set to the booting system, I will get video (monitor detection) errors - W98 sometimes (but not always) decides that the required monitor is no longer present and therefore reverts to a generic monitor which doesn't accept the selected refresh rates which usually drops the display down to 640x480 which doesn't always let you see the necessary buttons on the "display Settings" dialog so it is difficult to get back to the required resolution..... as so on and so on....... Similarly, the mouse is sometimes missed (much less frequent) - probably not a real big concern in a server room environment but a bit of a pain in the butt at home - I would expect that it is a combination of OmniCube _and_ Windows 98 interactions - it will be interesting to see if Windows ME has similar behavior. -the other problem is a bit more interesting. The OmniCube ocassionally locks up requiring major fiddling to get it working again. I think that occasionally W98 lockups, some power glitches and possibly some linux lockups occasionally lock up the switch as well - I can no longer change connections via keyboard or push button. Pulling the power cord doesn't do anything as the switch must be getting some back signal/power from the monitor and from the computers. Even switching the monitor off doesn't do anything (haven't tried unplugging the monitor from the wall yet). To get the switch back up and running I have to disconnect all of the video and monitor cables from the back of the box at which point the LED goes out and everything works as required when the connections are reconnected. Again not a real show stopper but definitely a pain when it happens. Keep up the good work. G^2 Guntis Glinavs Thanks for the kind words. As far as the booting problem, that's one that Belkin mentions may occur, and one that I've also experienced with other KVM switches, both manual and electronic. As far as the other, I wonder if power management is rearing its ugly head here. -----Original Message----- My last word on this topic: 17" "multimedia" monitor prices in Australia Philips $806.30 Notes All prices except Samsung obtained from a single source. [1] Hitachi have been undercutting Samsung prices for the last year in order to (successfully) gain market share. Samsung and Philips are the leading brands for market share. Note that the top 3 are very close to each other and the bottom 3 are close to each other. This was the first and second tier I referred to in my original email. Here in Australia, when quality regardless of price is the issue, the choice is between Philips, Sony and Mitsubishi. Where the task is less demanding, but longevity is still an issue, the choice is between Samsung, Hitachi and NEC. I find it interesting that the 2 brands you dismissed as worthy of consideration, Philips and Samsung, are at the top of their respective tiers. Below these top six, there is a host of cheap stuff that I wouldn't touch with a 10' barge pole. Jonathan Sturm [1] I couldn't find an online retail source of Samsung monitors. They are mostly sold in large volumes as part of systems. The retail source for the rest sells very close to retail price and we use them for deciding what we are going to charge locally. Since Hitachi are undercutting Samsung to gain market share, I think we can assume the Samsung price to be not too far off the mark. Well, those prices don't mean much to me without knowing the specific models and specifications. Most monitor manufacturers make at least three models--Good, Better, and Best--in each size range, and some make a half dozen or more. Within those ranges, the Best model can easily cost twice what the Good model does. It's obviously your business which brand of monitor you choose to buy. -----Original Message----- The problems you mention with WinMe sound familiar to me. When I switch my manual box and immediately start typing, without waiting till the screen stabilizes, I loose the keyboard. I do that often at work when I switch terminal sessions that I knock in my user-id and password in four sessions successively sometimes while the logon screen hasn't popped up yet. With my swicthbox I just have to wait for about a second. Maybe it helps to just wait a bit before moving the keyboard or typing on the mouse (well ...) >...recommending a $50 flashlight. looks a lot like the $20 light I use when working in dark spots (under a car, in the basement etc). Mine is a cheap plastic 3xAA battery thing with an uncomfortable double headband but it works good enough. And prices for such stuff (obviously) vary enough so you don't have to mention them in a book. Just like you don't mention prices of screwdrivers. Or did you? >...As far as monitor brands, ... Most 'lesser' brands don't make their own tubes but buy them in (MAG uses Hitachi and Sony tubes) and add electronics (bought in as well) and a casing. It is understandable that the tube makers use the best ones for their own brands and for their top customers (like Nokia, Eizo,...). One consideration, not mentioned yet, is that not all monitors from the same brand are equal. I know from experience that the smaller MAG screens (14..17") are much better then their bigger ones (19" and up). I wonder if that also happens to the top brands. For desktop LCD monitors another set of top brands will emerge. >Things are considerably less cluttered than they were... You know, you are on the brink of failing one of the daynotes requirements. That desk is unbearably tidy. -- -----Original Message----- Im really enjoying the seti thing. 2 points It would be nice to know what hardware people are using especially those with sub 10 hour units. My desktop is a PII 450 taking about 18 hours a unit my laptop is a 300 K6 with 64 MB and takes about 3 days! What does the program exercise? Ive started reading journal from mid 98 now (NT5 delayed maybe as late as mid 99 LOL) you refer to NT5 as MT5 may i ask what that meant? I can't speak for others, but I'm running SETI on half a dozen systems right now. The fastest is a Pentium III/600, which turns in times of just under 9 hours per unit. My Pentium III/550 does a work unit in about 11 hours. I have two Pentium II/300 systems that require about 14.5 hours each per work unit, and a Celeron/333 that requires a few minutes less than that. Most of the people with times in the 6 to 8 hour range are probably running fast Pentium IIIs and Athlons. As far as "MT5", I was under non-disclosure with Microsoft about NT5. Ordinarily, I strictly honor non-disclosure agreements and embargo dates. However, by that point, everyone (including Microsoft) was talking freely about NT5, so I found myself in the position of being under NDA, which technically forbade me from talking about what everyone else was talking about. But it didn't prevent me from talking about "MT5".
|
|
wpoison
Search [tips]
|
Wednesday, 30 August 2000 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] I did a little painting under high eaves yesterday, but we've decided to wait until we can make a Lowe's trip to do the rest. We were able to complete the eaves at the rear of the house over the deck, but the remainder would require working atop a ladder two stories up, and the idea of me doing that makes Barbara nervous. Not that I was crazy about the idea myself. Heights don't bother me, but standing atop an extension ladder means having to use a brush rather than a roller, which I prefer not to do. We're going to buy a six- or eight-foot aluminum extension pole that will allow me to reach the eaves while standing on a stepladder. That and some long-nap rollers. Between the two, I should be able to paint the eaves fairly easily. Actually, even if I need to work on the extension ladder rather than the stepladder, it'll still be easier. I can work half way up the ladder and pass the roller down to Barbara each time it needs to be reloaded. The problem was not the Belkin OmniCube. The problem was Windows Me. Just to make sure, I stripped anubis (the WinMe box) down to bare metal and reinstalled WinMe from scratch. Same problems. I then connected one of my Win98 test-bed systems to the Belkin OmniCube. It worked perfectly. I then stripped it down to bare metal and installed WinMe. After restarting it, I switched back and forth between the systems repeatedly and everything appeared to work normally, which made me wonder if perhaps there was something odd about the hardware on anubis. Fortunately, I then needed to do some real work on one of the other connected systems. When I finished that and switched back over to the new WinMe box, it was deader than King Tut. At this point, it was pretty clear that the problem was WinMe, so I stripped anubis down to bare metal and installed Windows 2000 Pro. Anubis is now named hathor, and so far it's been switching with nary a problem. The keyboard, video, and mouse on all three systems--WinNT4, Win98SE, and Win2KP--work perfectly. There's something about WinMe that the Belkin doesn't play nice with, but then WinMe has a lot of oddities. As far as I can see, there's absolutely nothing to recommend WinMe as an upgrade. Onvia.com is no longer on my list of recommended vendors. Last Tuesday, 8/22, I ordered an Epson printer and some paper from them. Both items showed as "In Stock" on the web page. A few minutes after I made the order, I received a confirming email from Onvia. That email included the information below, which also confirmed that both items were In Stock. Order Number: 465587 The paper showed up a day or two later, but not the printer. I didn't think much about that, because Onvia ships from multiple warehouses. But yesterday I started to wonder where the printer had gotten to, so I connected to the Onvia web site and went to the order tracking section. When I called up my order, it showed the printer as Back Ordered with no expected delivery date stated! There were lots of options on that page, such as reordering an item you'd ordered previously, but no option to cancel the order for a back-ordered item. There was an 800 number, though, so I called that. After half an hour or so on hold, I finally got through to a human. She apologized for the mistake and agreed to cancel the order. A moment later, she told me it'd been cancelled. "Are you absolutely sure?" I asked, "because I'm going to order this printer from another vendor and I don't want one showing up from Onvia." "Oh, yes," she said, "it's absolutely, positively cancelled." Being an experienced shopper, I asked her if she'd mind sending me a confirming email that stated that my order had been cancelled. No problem, she told me, and indeed that email showed up a few minutes later, at 7:42 p.m. As soon as it showed up, I turned around and ordered the printer and a spare set of ink cartridges from Outpost.com. But two hours after the original message, at 9:42 p.m., another message showed up from Onvia--the message that they should have sent a week earlier. That message apologized that the printer was back-ordered, and said that unless they heard otherwise from me, they'd ship it as soon as it arrived at their warehouse. I wasn't about to waste yet another half hour or so on the phone, so I replied by email, telling them that I'd already cancelled that order and that they'd damned well better not ship me a printer or place any charge for it on my credit card. So the upshot is that the cost of my wasted time has already far exceeded the entire cost of the printer, and it's all Onvia's fault. Chances are good that they'll ship me the printer once it shows up in their warehouses, and then I'll have to waste more of my time to get the printer returned and the charge removed from my credit card. In all fairness, I should be able to charge them my normal hourly billing rate for my time that they've wasted. Never again will I deal with Onvia, and I'd suggest that you avoid them as well. Barbara is off to the gym and grocery store, and then plans to spend the remainder of the morning volunteering at the library. When she returns from the library, we have to move all the furniture out of my mother's former downstairs area so that the carpet cleaning people can do their thing tomorrow morning. Getting a house ready to go on the market is a never-ending series of things that need to be done, but we're getting there. -----Original Message----- I'm sure you've seen the Registers article (or others)on Intel's recall of it's top of the line PIII 1.13GHz processor. It will be interesting to see how much longer Intel can make people fell they're getting more for their money buying Intel over AMD with all the problems they've had the last year. I remember you mentioning you displeasure with Intel when they released an essentially overclocked PIII 600 that required a higher voltage. It's deja vu all over again... Intel has certainly had a rough year, but I think they're past the worst of it. Frankly, I don't pay a lot of attention to the speed race up near 1000 MHz. Almost no one really needs a processor faster than the models that both Intel and AMD sell for $80 to $250. The 800+ MHz units from both Intel and AMD are grossly overpriced for the minor performance benefits they provide relative to the inexpensive 600 MHz to 750 Mhz units, and I really don't see much point to buying any of them. If I were building a personal system right now, I'd probably go with a $200 Pentium III/733. There's just no point to paying three to five times that much for a processor that provides minimal incremental speed. -----Original Message----- Just thought I'd chime in on the topic... Currently, I'm processing most of my work units on my Power Mac G3 (400 MHz PowerPC 750), which takes just under 7 hours on average to finish a work unit. I've also got a dual Celeron 433 system running NT4 that takes about 12 hours. I've only really been using these systems over the last couple of weeks, intermittently, once I joined the new group. Before that, I was mainly using my computers at work to finish work units. I had a couple of G3/400s, each taking about 7-8 hours, a Dell P-II 450 that took about 11 hours, and a Sun Enterprise 250 Server (dual 250 MHz UltraSparcs) that finished two work units every 8 hours. I was also using a Sparcstation 4 for a while, but since it took a few days to finish a work unit and was useless for anything else while doing it, I gave up on it. As for what the program exercises, I'm guessing from this data that it's mainly data-bound. From what I remember of FFTs from university, the x86 architecture has problems since it only has 8 floating point registers (vs 32 for each of the two RISC chips) hence it has to swap data between registers and (slow) memory more often. Also, the Apple and Sun systems have more L2 cache than most Wintel systems (1 MB for Apple, 1 or 2 MB for Sun -- I forget which), which probably helps with the large data sets SETI@home uses. Dan ------------------------------------------------------------------------- Thanks. You obviously know more about processor architecture than I do. I think you're correct about L2 cache size, though. The 512 KB half-speed L2 cache on Katmai-core Pentium IIIs seems to be just enough larger than the full speed 128 KB Celeron L2 cache and the 256 KB Coppermine-core Pentium III cache to give the Katmais a noticeable advantage. Perhaps the 1/2 MB L2 cache on your other machines helps further still. -----Original Message----- Hi Robert, I was curious to know if you had managed to get some Linux boxes working behind Win2k's ICS? I've been trying for the past couple of days and either I"m a complete dolt, or Redhat 6.2 doesn't get an IP from win2k's dhcp... If you have any info, greatly appreciated, otherwise.. I'll keep trying. Obviously I'm far from being knowledgeable in linux... it's just my little toy. -----Original Message----- Actually I figured it out.. seems that MSoft is insistent on changing the local adaptor's IP to 192.168.0.1 regardless of what it currently uses... which is awkward, but can be worked with, since everything is dynamic (Except for my linux boxes it seems).. Serves me right for skipping those message dialogs Microsoft pops up when you click OK. (It's not just an example, it actually changes the IP) But the real question you might be able to help me with... one I've been pondering for a while.. Using Win2k ICS, is it possible either with built-in features or with 3rd party software to map a port to a local machine? I.e. telnet connections to cableip@ port 5000 would be mapped to one of my linux boxes' ports? Sorry, but I've not used ICS at all on Windows 2000 (or on Win98, for that matter). Perhaps one of my readers knows the answer and can contact you directly. -----Original Message----- Bob - You may notice my rave review of my ATEN KVMS in my 08/29 Journal. It's a good piece of work, and has done well for me. I recommend you obtain an eval unit if you can,and check it out before you go too far out on a limb with Belkin. You might find you like it. Regards, JHR Oh, I'm not at all displeased with the Belkin OmniCube switch. The problem is Windows Me. The Belkin is rock-solid under Windows 98 SE, Windows NT 4, and Windows 2000 Professional.
|
|
wpoison
Search [tips]
|
Thursday, 31 August 2000 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] Our SETI@Home group continues to rack up work units, with about 7,000 units now complete. Congratulations to team member Wakeolda, who recently passed the 50 work unit milestone. If you haven't joined our SETI effort, please do. Here's how. Airborne showed up at about 1100 yesterday with the Epson Stylus Color 760 printer and the AIJ ink cartridges from Outpost.com. Barbara was in the kitchen making an early lunch and shouted back to me that there was an Airborne truck out front. I walked out the front door and mosied up to the truck, where the Airborne woman was standing looking into the side door of the truck and doing something with her clipboard. I waited patiently behind her until she finally picked up my boxes and turned around, whereupon she screamed. As others frequently comment, I'm really, really quiet for a big guy. Comes from that ninja training, I guess. Jerry Pournelle points out that if I have the same version of WinMe that he does--which it seems I do--it's a late RC rather than the RTM/Gold Code that I thought I had. That being the case, I withdraw all criticisms I made of WinMe until I have a chance to get the shipping version installed and running. My office is now clean. So clean, in fact, that some of my friends are giving me a hard time about it, afraid no doubt that their wives will see how neatly I routinely keep my office.
Barbara and I cleared the furniture out of my mother's former living area downstairs, preparing it for the carpet cleaning folks who are to arrive today. Here's what it looks like now, filthy carpet and all. We piled most of the stuff either in the kitchen or out in the garage. The sofa bed was so heavy that we decided it'd be easier to take the bathroom door off and stash the sofa in the bathroom, so that's what we did.
-----Original Message----- Hello Kevin... I saw your posting this morning on Robert Thompson's daynotes page and think that you might like to check out a piece of freeware called PortMapper [here] There are lots of interesting freeware things there, some are excellent, some are a little short on functionality for my needs, but this guy has some cool stuff. Regards, Kerry Liles - avid reader of Daynoters inbetween bouts of real work? ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -----Original Message----- Hi Robert: I found out about your wonderful web site from your posts to Jerry Pournelle's site. Keep up the good work! The reader who inquired about internet connection sharing and mapping a post to an internal system may be interested to know what I use, although I do not use it on W2K, but W98. I would bet it either works on W2K or the company (www.deerfield.com) will have a version out soon that does: I use Wingate Professional version 3 as an ICS program for DSL and internal firewall for my home LAN. It is (relatively) easy to configure the software to pass through any service to a specified machine sitting inside the network on a specified port. I would recommend he try it - I think they have a trial version available too. Wingate Pro is also good as a firewall for machines sitting behind it on the LAN, although you should make sure all the services are properly configured to prevent some accidental exposure. Also, it does not protect the host machine so I also run BlackIce v2 on that machine to help protect it from the internet (that does add one more thing to configure allow a connection into the network originating from the internet but can be done). This leads into what I am running the above setup on and asking you about your favorite uses for what is otherwise obsolete hardware: I am running the above on a AMD "586" 120 or 133 (I forget which) with only 36MB ram and Win98 first edition (with security updates). The performance as an ICS/firewall device is excellent, even with the windows overhead. I also use a 486 DX 33 with 32 MB ram running Win98 first edition (I was able to get it to ignore the too slow CPU during install) as a network connected back up device with similar functionality to the Quantum Snap servers. I bought a 20 Gig Maxtor hard drive when they first became cheap for this unit and it has perfectly fine (although not the speediest) performance backing up other machines over the LAN and was a lot cheaper than the $700+ price I've seen on the network plug and go disk servers! I am running a Packard Bell 386 SX with 4 MB Ram as a fax receiving machine (using BitFax Pro software) on WFW 3.11. When faxes come in I can simply open up the drive from my main machine and click on the new fax file to view from my main machine. I read something about you using a 386SX for an automated attendant. Would you be willing to explain that setup and software? Any other suggestions for productive uses for old computers is always welcome. Thank you again. Sincerely, Bruce Edwards Thanks for the kind words. Obviously, there are any number of things you can do with older hardware, many of which involve running Linux on it. As far as the automated attendant, I'm running a telephony card called a BigmOuth, from Talking Technologies, Inc., which has unfortunately apparently gone bankrupt. I chose that card because it was the only one capable of doing supervised (or "smart") transfers. That means that when someone presses a key to transfer a call to me, the BigmOuth watches the progress of that transfer, and can take different actions depending on whether the transfer is busy, ring-no-answer, etc. No other consumer-priced card on the market could do that back when I first installed this system years ago, and I suspect none can now. If I'd known that TTI was going bankrupt, I'd have bought a spare card, but I didn't find out about it until well after the fact. If this card ever dies, I suppose I'll have to buy a high-end Dialogic or something. -----Original Message----- Robert Last week I needed a couple of ink cartridges. I looked at Onvia.com's web site. I had never been there before. I found it hard to navigate and the item descriptions were poor. I went to buy.com and ordered the cartridges on Friday. I received them yesterday (Tuesday). Mike Boyle I never noticed that. When I'm going to buy something, I generally research it thoroughly at the manufacturer web site, and then buy by manufacturer part number. I did notice that Onvia for some reason tends to use Canadian rather than US part numbers, which makes comparing more difficult. I don't know why they do that, except that they have both US and Canadian operations. -----Original Message----- Robert, You mention how the big mail-order vendors swap parts on like systems. I've had this experience in spades with Dell, and it can be a real support issue. Also, I've experienced a ridiculous amount of failures on Dells in the last 10 months. I'm now considering a change in vendors for end-user and developer desktop and laptop systems with a Wintel platform. Any thoughts? I'm leery of building my own, because no one on my staff has the time, and adding staff seems to run counter to cost-effectiveness. On KVM switches: I'm using Belkin OmniCube in my current environment. I occasionally have problems like those you describe on NT 4.0 SP5 boxes. I believe it is due to the lack of a continuous power signal sent by the Belkin box. In my previous position we used KVM switches on Wintel boxes that were app servers for a real-time production environment. We used Black Box KVM switches which worked flawlessly under all conditions. They sent a continuous power signal to all connected boxes, avoiding the "dropping" of a mouse or keyboard by the OS. They also had a "reset" feature which fixed the "jumping mouse" problem instantly. Of course, they cost about 5 times as much as Belkin, which is why I'm not using them now. The occasional problems are no big deal in my current environment, whereas saving money is a big deal indeed. --Dave Dave Pierce I'm surprised that you're having problems with Dell units. Over the years, I'd come to regard them as the most reliable of the direct resellers. If you don't want to build your own, you might consider contracting with a local screwdriver shop to build units to your specifications. You supply the parts (or provide detailed specs for what you want in your systems), and they assemble and test the units. Many local shops will provide assembly services for something like $50 to $75 per unit, which still ends up costing you less than buying at retail and gives you a system with much better components. As far as the KVM switch, I've had no problems at all with the Belkin OmniCube on systems running Windows NT 4, Windows 98, and Windows 2000. It's only on systems running Windows Me that I have problems. -----Original Message----- Was reading that you are using rollers to paint....wondered why you didn't pop down to the local rental place to get a spray painter. Although the cost is about $100, the time saved is pretty impressive, and coverage is better, IMHO. Last time I did the house, I didn't have to mask anything, as the trim around the windows was wide enough (3 1/2 ") that I could carefully spray around the windows without getting any overspray on the windows. It was also good for under the eaves, which I'd think is very time-consuming. After the big spray on, I took the roller to the trim. Secret trick: only roller the front of the trim, no need to do the trim sides. I also used the roller (on a long pole) to get the front of the rain gutters. Had to do a bit of hand-brushing along the end-of-house-eave trim (front edge of the roof trim at the gable ends of the house). I only have a one-story house, the gable ends are 16' (approx) high, but was able to paint from the ground with an extension (6 ', I think) on the sprayer. Eye goggles and a hat are musts, though. I don't think that I'd want to paint the house by roller....even for a rookie, it's not hard to do a good job around windows and doors. Although that depends on the trim....I did do some masking around the front wood door and windows there. I suspect that you may have considered this, but post this for the other lurkers too. Regards....Rick Hellewell Nah, I've got my own spray painter, but Barbara won't let me use it. Actually, we're using brushes and rollers. Barbara does the detail work (she can paint the separators between panes with a 4" brush without getting paint on the glass) and I use brush and roller. I cut in where the roller can't reach with a trim brush and then use the roller to cover the wider areas. It actually goes pretty fast. -----Original Message----- Robert, If you don't know the answer, perhaps one of your readers could help me. I'm running a red hat 6.2 and KDE. I'm having a terrible time double clicking my mouse fast enough in netscape to highlight text in the URL line. I can double click fast enough only rarely. My wife can't do it at all and my 8 year old (with stunning Nintendo-trained fast twitch muscles) can only do it about 1/2 the time. I've hunted in the documentation, looked in deja news, posted questions on usenet and I can't find a way to slow down the mouse click speed. Perhaps someone with a wizard hat can favor me with knowledge. Mark Huth No idea, but I'm sure one of my readers will know the answer.
|
|
wpoison
Search [tips]
|
Friday, 1 September 2000 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] Good news about PC Hardware in a Nutshell. I spoke yesterday with a woman at O'Reilly who does marketing/promotion, and she tells me that the "sell-in" (pre-orders) for that book are quite good. When I mentioned them to my agent, he also sounded excited. I won't mention specific numbers, because O'Reilly probably considers them confidential, but they're about 50% higher than the number that they would have considered "good". Better still, they don't have all figures in yet, because the O'Reilly sales folks are still out there beating the bushes for additional orders for "fall titles". So the numbers may end up another 10% to 25% higher than they are right now by the time the book actually hits the warehouse. That, incidentally, is scheduled for 22 September, which means the book should be available in bookstores before the end of this month. Of course, the sell-in numbers don't really count for much. They simply represent orders made by booksellers and distributors. What really counts are the "sell-through" numbers, which represent sales to actual people. But we won't have sell-through numbers (re-orders) for a couple of months, so the sell-in numbers are the best we have to go by at the moment, and they're looking pretty good. After putting in 18 months of grueling work on this book, we're keeping our fingers crossed that the numbers start high and remain high. So now my priorities are getting the stuff done that will help the book move into the channel and sell a lot of copies. Right now, that means doing an article or two for the O'Reilly web site and getting the web site for the book, hardwareguys.com, up and running. That means that the posts I do here are likely to be shorter for the next month or so, because much of my available time will be allocated to getting the book's web site fleshed out by the time the book hits the stores. Our SETI@Home group continues to rack up work units. Congratulations to team members BFMersion and Peter Bruno, who recently passed the 100 work unit milestone. If you haven't joined our SETI effort, please do. Here's how. The KVM problem may be the mouse, or more specifically that the IntelliMouse Explorer is not supported by the Belkin OmniCube KVM switch. I switched that mouse for a standard mechanical IntelliMouse, and everything appears to be working properly. I'm not sure why the Belkin (or the systems) should care whether there's an IntelliMouse or an IntelliMouse with IntelliEye connected, but it appears they do. At any rate, with the mechanical mouse connected, the Belkin OmniCube appears to be 100% stable with all operating systems, including WinMe. I've had numerous reports from readers of similar problems, and it appears that simply changing the mouse may be the solution to those. As I was working at my desk yesterday, I accidentally bumped hathor with my knee. The front bezel popped off, which turned out to be a fortunate thing. In retrospect, I last cleaned hathor (which was then named kerby) about six months ago. When I was rearranging my office the other day, I did a quick vacuum of the front and back panels, which didn't look bad at all. So I assumed that the machine had remained fairly clean, sitting as it was on top of the desk rather than on the floor. That obviously turned out not to be the case. The combination of dog hair and pipe smoking generates a real mess at the air inlets on a PC.
No real harm done, though. I fired up Barbara's Ferrari™ vacuum cleaner, which sucked all the dust and junk right off. I think it may have also sucked some chips out through the holes in the front panel, but everything appears to be working okay. We got the Epson Stylus Color 760 inkjet printer unboxed and connected yesterday. It has both parallel and USB interfaces. I was going to install it back in Barbara's office, but she runs NT4, has only one printer port, and already has a LaserJet printer connected to it. Also, she didn't really have anywhere convenient to install the printer. So it ended up in my office, connected to my Windows 98 SE test bed via USB. I dumped a couple of test prints onto standard Xerox paper. I must say they look surprisingly good given that they're on cheap paper and at only 72 dpi resolution. Given that the Olympus D400-Z does 1280X960 resolution, I plan to print most images from it at 240 dpi, which yields a print 4" X 5.33" (10.2 X 13.5 cm), just shorter than a standard 4X6" print. I may also try 192 dpi, which'd yield a print 5X6.7, just under the standard 5X7. I'll probably try using 120 dpi, which'd yield more-or-less a standard 8X10, although I think that may be stretching it. We're still getting the house ready to go on the market. The carpet cleaning folks showed up yesterday, and the carpet in the downstairs area now looks a lot better than it did before they cleaned it. This morning, we're off to Lowe's Home Improvement Warehouse to get some supplies--long-nap paint rollers, a roller handle extension, and so on. We had a comprehensive list of the stuff we needed, added to as we thought about it. I say "had". Barbara made the mistake of leaving it on her end table yesterday so that we wouldn't forget it when we left for Lowe's this morning. Unfortunately, Malcolm was in a pillaging mood yesterday, spending a lot his time perusing our end tables. He found the list and ate it. Actually, he did leave a small corner unconsumed, enough for Barbara to identify what he'd eaten. That's fortunate, because otherwise we'd have been scurrying around the house looking for it. -----Original Message----- Mark - I swam back through the KDE 1.1.2 version I have installed with Caldera eDesktop - this is the same version as 6.2 RH, I belive... Yup, confirmed on a RH RPM mirror site... The X-Server is also about the same, 3.3.6, though possibly a different sub-minor rev. All this is in preface to saying - nope, haven't found a way to adjust the doubleclick interval in KDE 1.1.2. No joy at the KDE site, on Deja, on the KDE usenet group, or in the various configuration utilities I know of. Hmmm. I have posted a question on the usenet group - we'll see what comes of it, because that is a useful thing to be able to adjust. I think I caught a hint that KDE 2.0 will have something along those lines, but can't be sure, and I don't have access to the 2.0Pre that installed with Caldera LTP at the moment. I'll let you know if anything more turns up. In the meantime, click once and hold, then drag to mark a block, rather than double-clicking to grab the whole thing in one whack. In other news, the Galeon browser, which uses the Mozilla Gecko engine with a clean (not quite ready for prime-time yet) interface, has the following behaviour: doubleclick on in a URL on the address bar, and the WORD is highlighted, so if you were at www.orbdesigns.com and you wanted to go over to Bob's place, you can double-click to just select orbdesigns and type/replace w/ttgnet. Very handy and something I've been wanting for a very, very long time. Take care, Brian -- -----Original Message----- Yes, I've also had good experience with Dell over the years. However, I now work for a startup which has only existed for 18 months. We own approximately 40 Dell boxes. We've had four complete failures, and 10 component failures. Definitely sub-optimal. Your "screwdriver shop" idea is exactly what one of my staff suggested. I think he reads your site too. Good thinking, Dennis... :) Wow. That's an extremely high failure rate, probably ten times higher than I'd have expected from Dell. No wonder you're looking at alternatives.
|
|
wpoison
Search [tips]
|
Saturday, 2 September 2000 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] Our SETI@Home group continues to make progress, with about 7,500 work units now complete. Congratulations to team members Jim Kershner, who recently passed the 1,000 work unit milestone, and Dave Browning, who recently passed the 50 work unit milestone. If you haven't joined our SETI effort, please do. Here's how. Barbara is headed over to her parents' house today, but before she leaves she's moving stuff back into my mother's former living area downstairs. I helped haul in the sofa-bed (which is apparently constructed of solid oak with steel rebar) and the chest of drawers, which really is solid oak, I think. More interesting stuff showed up yesterday, including an 80 GB Maxtor IDE drive. Actually, it's about 81,000,000,000 bytes, which translates to a true 75+ GB, but all hard drive manufacturers long ago abandoned specifying true MB and GB in favor of millions and billions of bytes. At any rate, it's the largest disk drive now available of any sort, including SCSI. In fact, it's within striking distance of the absolute ATA limit on hard drive size. If Maxtor wanted to, I suppose they could build a full height drive that ran up against that limit. When I have a spare hour or two, I plan to install the Maxtor in a test-bed system and benchmark it. The one widely-reported downside to the 80 GB Maxtor drive is that it runs at 5,400 RPM rather than 7,200, a step that Maxtor likely took to reduce heat, noise and vibration. Obviously, that slower rotation rate penalizes average access time slightly (a 5,400 RPM drive has rotational latency of about 5.56 ms versus about 4.17 ms for a 7,200 RPM drive), but the Maxtor has a fast enough seek time to mostly make up for that 1.39 ms penalty, particularly for sequential transfers. Given its very high data density (and accordingly high DTR), I expect the Maxtor to provide quite high performance despite the 5400 RPM spindle speed. But we'll see. -----Original Message----- Bruce, When you said today "I dumped a couple of test prints onto standard Xerox paper", are you actually using Xerox-brand photocopier paper or are you just trying to turn Xerox into a public domain word like happened to aspirin? As for your KVM problem with the IntelliMouse Explorer, is this one of those new mouse that glow red and could the power required for the LED be related to your problem? (And is mouse the correct usage here?) Just wondering (in both cases). Thanks for your column and I eagerly await being able to purchase your new book. It will definitely be a useful reference volume. No, it's actual Xerox paper. It says so right on the label. We buy it at the local Office Depot for something like $5 a ream. Yes, the mouse is one of the new "red light" models, as Pournelle calls them. I can't imagine that the power needed to run the LED is a problem, but perhaps you're right. As far as the book, it should hit the warehouses on the 22nd, and be available in bookstores within a couple days after that. -----Original Message----- Bob - In future iterations of your PC hardware books, you may wish to venture into a chapter on the differences in effects and results of Linux vs. Windows OSs with identical hardware. I recently discovered a couple of them. (full report here.) 1. HCLs from at least one Linux SW vendor (Mandrake) are NOT an indication of compatibility or actual support. 2. Mandrake Linux somehow manages to override or bypass the BIOS setting that disables all power management. I suspect there are other even more subtle but serious differences that I have yet to have the pleasure of experiencing. <BG> Regards, JHR We actually talked about doing that, but the problem is that I write only about things I know, and I don't know Linux. Still, we'll see. -----Original Message----- Bob, On the topic of Dell quality... one minor gripe. The Dell Dimension XPS T600r that I picked up earlier this year for use as my at-work workstation arrived with the mouse port jammed. When the Dell techs assembled the mobo into the case, one of the case shielding fingers went into the open space between the PS/2 port shield and connector body, which prevented the mouse connector from plugging in at all. I fixed it myself. This tells me that (a) they forced the mobo into place, often potentially bad news, and (b) they couldn't have tested the system post-assembly. I had forgotten that I regretted not following your wonderful build-it-yourself advice that you gifted me with prior to the purchase. Hmmm. Yes. I know. Whack, once, with a self-applied LART. <g> .b bilbrey@orbdesigns.com Well, everyone screws up from time to time, but I'm hearing enough complaints about Dell recently to make me wonder if they've really lost it. The sad fact is that although you'll always get a better computer by choosing the parts yourself and building it yourself, few businesses have the time or staff to do that. -----Original Message----- I've actually seen this pattern before. Compaq went through it and so did GW2K. I think what happens is they reach a critical mass of size and reputation to where they become the default choice for their target market. At that point some MBA figures out that people buy their stuff regardless, so they can increase profits by millions by getting rid of the quality components and quality processes that got them there. I always try to vote with my feet when that happens, but sadly the companies seldom see serious consequences for it. I believe it's just Dell's turn now. They probably hired a Wharton graduate last year... You may be right. I hope not, though. -----Original Message----- Robert: I showed my wife your picture of the former Kirby with it's face plate off. She sentenced me to computer cleaning this weekend. Do you think naming your computer after a vacuum cleaner might have created a reality anomaly? On the Dell failure rate, has the power system been checked for low power condition? We experienced a high failure rate in Compaq computer workstations, which suddenly ended when we insured that all equipment was plugged into the proper protected outlets (UPS including diesel generators if the batteries run out). The contract installers had not been informed of the letter coding on the outlets. Heads are expected to roll at any moment. Jeff Timm Who enjoys and applauds the efforts of those doing things so I won't have to. Actually, kerby was named for one of Barbara's (large) collection of stuffed bears. Kerby (the bear) is about 20 years old now. Although Barbara has many antique stuffed bears, Kerby is the oldest of the ones she's bought new since she started collecting them.
|
|
wpoison
Search [tips]
|
[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] Our SETI@Home group continues to make progress, with nearly 7,700 work units now complete. Congratulations to team members Shawn, who recently passed the 1,000 work unit milestone; Tom Syroid, who recently passed the 250 work unit milestone; Paul Robichaux, who recently passed the 100 work unit milestone; and Keith Soltys, who recently passed the 50 work unit milestone. If you haven't joined our SETI effort, please do. Here's how. I want to like Sun StarOffice, I really do. After all, if it worked as advertised, it'd be a reasonable replacement for most of the components of Microsoft Office, and it's free. Unfortunately, StarOffice is a hard product to love, and an easy one to hate. I installed StarOffice 5.0 for Windows when it was first released. I tried to use it, but it was so slow and the interface was so terrible that I eventually gave up. I downloaded 5.1 when it was released, and found few improvements. Slow, clunky, and ugly pretty much described those versions. Then, a month or so ago, I downloaded StarOffice 5.2 (for Windows and Linux), hoping against hope that they'd have done something to improve the performance and the interface. Alas, it appears that I'm destined never to know. I had a few spare minutes yesterday, so I decided to install StarOffice 5.2 on kiwi, my main Windows NT 4 Workstation system. The installation proceeded normally enough, and I accepted all default options. When installation completed, I was faced with a dialog that told me I needed to restart the system before using StarOffice. Fair enough. There were two buttons, one of which allowed me to re-login and the other of which quit setup. Not trusting any button that wanted to quit Setup, I chose the re-login button. That one shut down and restarted Windows (although not the PC itself.) Okay, perhaps I'd misunderstood the last dialog. Maybe all it wanted me to do was logout and then login, which was what I'd just done. So I started StarOffice, which displayed a splash screen for several seconds, which cleared to display a mangled desktop. My mouse didn't work, and I couldn't even Ctrl-Alt-Delete to bring up Task Manager and kill the process. No option but to power-reset the machine, which is annoying. Once the machine re-started, I un-installed StarOffice and rebooted yet again. After deleting left-over files and directories and restarting again just in case, I fired up StarOffice Setup and tried again. Once again, Setup proceeded normally, and I chose all default options. This time, when I got to the final screen, I told Setup to quit. It did so and restarted the system. Aha, I thought. All I'd needed was a full reboot. So I fired up StarOffice, which again blew up and locked the system up solid. After yet another power reset, I again un-installed StarOffice, and again restarted the system. This time, rather than just deleted the C:\Office52 folder, I decided to check the registry with RegEdit. Sure enough, StarOffice had left all kinds of garbage--dozens of registry keys--scattered all throughout my registry. So much for their uninstall. Wanting to be fair, I again tried installing StarOffice 5.2, this time on a cleanly formatted test-bed system running Windows NT 4 Workstation. Short story--same mess. Perhaps it's unfair of me to expect a free product to at least install successfully, but I don't think so. If this is the best Sun can do, Microsoft has nothing at all to worry about. I'll probably try StarOffice again, but not until it reaches at least version 6.0. For now, at least, this is a dog of a product and I won't waste any more time on it. And I'm wondering what crappy DLLs it installed to replace perfectly good ones in my main Windows directories. So I suppose I'd better re-install SP6a before I do much more on this, my main system. Well, Barbara wants to work on getting the downstairs packed up, so I'm going to help her while I'm doing laundry. We hope to get the entire downstairs kitchen cleaned out and packed up today. -----Original Message----- Hi Robert, Thought I'd add my 2 cents in on the Dell issue. My company standardized on Dell last December, since then we have bought about 220 Optiplex desktops and 30 notebooks. We've had 1 monitor fail (immediately) which was replaced and on site by 9am the next morning.This could have been a shipping problem as the box was damaged. Other than that, so far no problems. In fact we find that we have had no service issues and significantly reduced user issues wherever we have replaced one of our former systems with a Dell desktop. I will however agree that the Dimension series seems to have more QA problems than the Optiplex series. If I was to buy a Dell for home I think I would forgo the Dimension series and buy the Optiplex series even though they are meant for corporate desktops. Actually I'd probably continue to simply build my own for home and am looking forward to buying your hardware guide. (For two reasons, one it's an O'reilly book, and two, I won't need a yardstick to measure the width) Good point. Although Dell and some other direct PC sellers have two product lines that they try to differentiate by such things as bundled software and peripherals, the truth is that there's often a qualitative difference between their less-expensive "home" systems and their more-expensive "corporate" systems. Of course, they can't very well admit that their home/SOHO boxes are of lower quality than the flagship corporate products, but I suspect you're right that there is a real difference. And even their more expensive product lines aren't as good as home-built systems, which you obviously already know. -----Original Message----- Robert, I was poking around the HardwareGuys.com web site and noticed that the mailing address for the HardwareGuys.com webmaster at the bottom of the home page and in other locations on the site is missing the 'w'. Presumably this is unintentional. Like many of your readers, I'm looking forward to seeing the book on the shelves of my local bookstore. Also, is Jerry Pournelle a collaborator on the book? I was under the impression that he was, but I only saw you and your wife listed as authors. I'd already seen that one, but thanks for pointing it out. I'm working on the hardwareguys.com web site locally, and haven't published it up to the server for quite a while. Once it goes live, I'll be updating it frequently. As far as PC Hardware in a Nutshell, that book is by my wife and me. It'll run 500 pages or so. Pournelle and I are working on an expanded version that'll run twice or more that size (and probably cost twice as much). That book will be called PC Hardware: The Definitive Guide. -----Original Message----- I just joined your SETI group. I've been running the client for almost a year now, but on one of my older computers. I just added it to my main machine (a PII 300), which has boosted my units a bit - I just hit 50. I don't keep a daynotes journal (no time) but very much enjoy reading yours. Best -- Thanks. If my own experience on a couple of my PII/300's is anything to judge by, you'll be cranking out a work unit about every 14 hours or so. -----Original Message----- > but all hard drive manufacturers long ago abandoned specifying true MB and GB in favor of millions and billions of bytes Unfortunately, the hard drive manfacturers are right. "Mega" means one million. "Giga" means one billion. A kilometer is not 1024 meters! The problem is that this is a 2.4% error. But if we use 1024*1024 as "mega", it's a 4.9% error. Giga is a 7.5% error. Tera is a 10% error. It just gets worse. There are accepted prefixes for 1024, 1048576, etc.. kibi (Ki) and Mebi (Mi). and so on. They're just not popular. But much worse I think is the bit/byte confusion, where some companies list BITS (in RAM chips for example) because it makes the compnent or value sound bigger. Even today, I don't know the real definition of "baud". It's not bits per second.. what is it? Thanks for the site, I read it every day. :-) -Steve No, the hard disk manufacturers are not right. That's why they always put a small asterisk next to their capacity numbers. Kilo, mega, and giga have defined and universally accepted meanings in the computer world, as a binary thousand (2^10), a binary million (2^20), and a binary billion (2^30). Any time anyone refers to a binary value, such as the size of memory or a hard disk, the correct definition of the prefix is by the binary value. There's no confusion there, except what the marketing folks have intentionally introduced. The bastardized prefixes you mention are not accepted in any sense, and simply represent the attempts of a couple of standards bodies to solve a non-existent problem. Kibi, mebi, etc. were proposed by the IEC in, I believe, 1997, and subsequently adopted by NIST, but have since gone nowhere. Nor are they likely to. As computer-literate as my readership is, I'd be willing to bet that not even 10% have even heard of the new prefixes. Baud measures state changes/second. The original 300 baud modems encoded one bit per state change, and so were also 300 bps. At higher speeds, multiple bits are encoded per baud, so that 2,400 bps modems, for example, actually ran at 600 baud, encoding four bits per baud, and 9,600 bps modems at 2,400 baud, also encoding four bits per baud. The 14,400 modems also modulated at 2,400 baud, but encoded six bits per baud. And so on. [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] |