|
Week of 30 August 1999 Sunday, 05 September 1999 09:42 A (mostly) daily journal of the trials, tribulations, and random observations of Robert Bruce Thompson, a writer of computer books. |
|
|
Monday, 30 August 1999 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] I see over on Ars Technica that Unisys, that pathetic excuse for a computer company, is now trying to extract a $5,000 license fee from web sites that have .gif images available for download. Not all web sites, mind you, just intranet web sites and small public web sites like this one. They plan to negotiate individual licenses with larger web sites. This is actually round two of this money grab. Some years ago, Unisys went after software vendors whose products created or displayed .gif images, forcing them to buy a license. So any software you might use to create or view .gifs is almost certainly licensed already, and there is no need to license images that were created by a licensed programs. But Unisys thinks I should take a better-safe-than-sorry approach by paying them a $5,000 license fee just in case an unlicensed program sneaks in. Yeah, right. They're pathetic. Apparently, they can't compete in the computer industry, so now they're looking for new sources of revenue. Obviously, they've now run out of software vendors to club over the head with their ridiculous patent (their LZW patent is based on the public-domain Lempel-Ziv algorithm), so now they're going after individual web sites. That's disgusting. It's particularly disgusting because .gif is an obsolete, crappy image format. There are much better formats, like .png, that are in the public domain. Just as an experiment, I converted a .gif format image of my tartan and converted it to .png. Here it is in .gif format, which occupies 20 KB. Here it is in the lossless .png format, which occupies 2 KB. Now, why doesn't IE5 have built-in .png support? Even Netscape Navigator, which hasn't been updated in two years, supports .png files natively. IE5 just offers to download the graphic. Geez. * * * * * This from Svenson Sjon [sjon@svenson.com]:
How much? Not enough. How? That depends on what you're writing. With fiction, you generally write the book first and then your agent shops it around to various publishers or puts it up for bid. The publisher then pays an advance, which varies according to the genre and how well-known the author is. An obscure author who writes a mystery may get an advance of a few thousand dollars. Patricia Cornwell got a $24 million advance for two books. For hardbacks, the fiction author is paid a royalty (typically 15%) on the list price. That means that each $30 novel earns the author $4.50. Once enough royalties have been earned to pay back the advance (called "earning out"), the author begins to receive royalty checks. Once the book goes to paperback, the author gets a much smaller royalty per copy, sometimes as little as $0.10 or $0.20, but many copies are sold. With computer books, the process is similar, but advances are normally paid while the book is being written. Advances for computer books are typically in the $5,000 to $20,000 range, depending on the size of the book, the author's reputation, etc., and are normally paid in chunks, e.g. 25% on contract signing, 25% on submission of the first chapter or two, 25% at 50% completion, and 25% at 100% completion. Royalties are similar to fiction, but are generally at a lower percentage level (typically 10%), and are calculated on net rather than gross. That is, a computer book that lists for $40 might be sold by the publisher to the distributor for $20. The 10% royalty is calculated on that net, yielding a royalty of $2.00 to the author. Foreign sales are usually at 50% the royalty level of domestic sales, so a nominal 10% royalty rate often turns into a net royalty rate of 7% or 8%. Very high-volume books (e.g. the Dummies series) often pay very low percentage royalties (on the close order of 3%), and an author of one of those often earns as little as $0.25 per copy. * * * * * This from bdenman [bdenman@FTC-I.NET]:
Software, drivers, and so on can certainly cause system stability problems, but hardware is fundamental. Intel didn't abandon Socket 7 for competitive reasons. They did so because Socket 7 had reached its limits of reliability. AMD and the third-party chipset makers have pushed Socket 7 even further. They can call it Super7 or whatever they want, but the fact is that Socket 7 simply isn't good enough for reliable operation at the speeds today's processors run. AMD acknowledged this by going to Slot A for the Athlon. If AMD could have licensed GTL+ in either Slot 1 or Socket 370 form at a reasonable price, you can bet they would have done so. A GTL+ version of the K6-III would have been competitive with the Celeron, but a K6-III hampered by Socket 7 cannot compete on stability. Nor are the Via, SiS, and ALi chipsets as robust as Intel chipsets. But Intel abandoned Socket 7 chipsets with the 430TX, which is no longer competitive on features. As far as using brand-name RAM and power supplies, all I can tell you is that they make a difference based on my experience over twenty years with hundreds of systems. We're seeing some effects from Dennis now. It's very cool this morning, and the winds are gusting up to about 30 MPH. No rain yet, but we're hoping to get at least an inch or two from Dennis.
|
|
|
Tuesday, 31 August 1999 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] I'm usually up and working at my PC by 7:30 a.m. or so. But I slept in this morning until a few minutes ago and have some things I need to get done this morning, so the usual ~ 9:00 a.m. update will be delayed until later this morning or this afternoon. Lots of interesting mail today. We begin with a warning about Windows 2000 Professional and DHCP. It all began when I was reading Dr. Keyboard's diary this morning, when he reported some problems connecting to his server. I sent him the following: IIS, for some bizarre reason, defaults to using default.htm as the home page. But it's easy enough to change that to index.htm, and in fact IIS can be configured to support multiple home page filenames. When I set up IIS, I normally configure it to respond to index.html, index.htm, default.htm, and default.html in that order. As far as your problems with W2KP, do you have it pointed to your ISP's DNS server or a local one? It should be the former. Also, when you configure your dialup connection, make sure to leave the Windows NT domain name blank rather than putting in the name of your local Windows NT domain. To which he replied:
To which I replied: Well, that's truly bizarre. I don't know if you're using DHCP or not, but if you are you might want to consider assigning static IP addresses to all of your machines instead. I'm beginning to think that Microsoft's NT5 version of DHCP means "Dynamic Host Confusion Protocol" instead of "Dynamic Host Configuration Protocol". Just a couple of months ago, I brought up a W2KP workstation configured as a DHCP client. It proceeded to assign itself the same IP address as another workstation on my network, and never said a word about it. I'd have missed the problem entirely except that the other workstation happens to sit on my credenza and beeped at me. When I moved the mouse to unblank the screen, there was a warning box telling me that another host on the network had taken the IP address that it was already using. That IP address (mis)assignment was truly gratuitous, as the DHCP scope on NT4 DHCP Server had a block of addresses that occupied most of a C-block, and only a dozen or so of those addresses were already in use. The NT5 DHCP client apparently just has to pick one that's in use instead of one that's not. I've since repeated the experiment, and it did the same thing to me, but this time with a different workstation. This was all with Beta 3, and I'd have hoped that they'd have fixed the problem in RC1, but apparently not. * * * * * This from Joshua D. Boyd [jdboyd@cs.millersv.edu]:
I understand. From my recollection of what you said earlier, it seems that your motherboard supports at most a 200 MHz Pentium. That'd mean a 66 MHz bus and a 3.0X multiplier. I'm not familiar enough with the AMD K6-* CPUs to know for sure, but many Socket 7 CPUs translate CPU multipliers. That is, rather than use the nominal multiplier value, they substitute some other multiplier. So, for example, setting the motherboard jumper to a 2.5X multiplier may in fact translate to a 5X multiplier with a given processor. I don't know how AMD handles this, but it's something worth checking out before you buy one. * * * * * This from Gary M. Berg [Gary_Berg@ibm.net]:
Yep, you're right. I tried it on another system with IE5, and it worked fine. But it doesn't work on my main workstation, and I can't figure out why. I've checked everything I can think of in IE5 options, file associations, etc. and there's nothing apparent to account for the problem. When I click on the link to the .png graphic in my copy of IE5, it brings up the dialog box asking if I want to download the file or open it. If I choose to open it, IE5 invokes IrfanView (my graphics viewer) and displays the graphic properly. Obviously, the fact that I installed IrfanView on this system accounts for the problem, but I can't figure out how to turn off IrfanView as the default viewer for .png graphics. I've gone into IrfanView options and unchecked the .png box, but that doesn't change the default behavior. * * * * * This from Daniel Seto [dkseto@email.com]:
Well, perhaps, but all that really proves is that chess is trivial in a mathematical sense. Well, perhaps not yet trivial in the same sense that tic-tac-toe is mathematically trivial, but certainly on the verge of becoming so. Once we have machines fast enough to calculate all possible end positions, chess is no longer a game. Compare that to the vastly more complex contract bridge, which machines are not likely to play well for many years to come. I don't think that we humans are in any great danger of being overtaken by computers. * * * * * This from Frank McPherson [frank@fmcpherson.com]:
Yes, from your description I'd guess that you almost certainly have power problems. The fact that the same problem occurs with a different PC and monitor in a different location makes it a near certainty. I've seen similar problems over the years. The most extreme one was a PC that would literally reboot itself spontaneously many times a day. We finally tracked that one down and found that a postage meter in a room several rooms away was causing the problem. Every time someone ran the postage meter, the PC rebooted. To absolutely verify the problem, the first thing I'd try is plugging the PC into a different circuit. It's quite possible that both the first and second PCs are connected to the same circuit, which would account for them both behaving similarly. The solution may be as simple as plugging the PCs into different receptacles in the same room. Just getting them off the original circuit may make the problem go away. But it may also be that your electrical service is simply inadequate to provide smooth power under all conditions. This is pretty common in apartment buildings, where sometimes the voltage drops below nominal every time someone fires up the elevator. If that's the case, you need a UPS that has a "boost" function that uses the battery to maintain mains power at nominal during sags. * * * * * This from Tom Syroid [tsyroid@home.com] regarding a new Microsoft e-book initiative:
Thanks. I wish Microsoft success on this one. It would be nice to have a dominant e-book standard in place. To quote Churchill, "this is not the beginning of the end. It is, perhaps, the end of the beginning." I think the article was pessimistic in one sense. It mentions 2020, but I don't think it'll be anywhere near that long before e-books become a major factor. My guess is that e-books will become common by 2005 and dominant by 2010. What's holding them back now is not technical issues, but publishers protecting their interests. In the same way that the music industry uses contracts to ensure that they will have a major share of electronic music distribution, publishers will attempt to ensure that they keep their hooks into content creators. In the long run, it won't work for either of them. The music industry is in a bad position, because all they provide is distribution, and their distribution channels are rapidly becoming obsolete. Book publishers are in a somewhat better position, because they provide layout, design, and editorial services as well as distribution. But their main value-add at this point is distribution, and that too is going away. I do know that my next book contract will reserve all electronic distribution rights to me, period. I attempted to do that on my last couple of book contracts, but the publisher wasn't willing to give them up. * * * * * This followup from Tom Syroid [tsyroid@home.com]:
* * * * * This from Robert Rudzki [rasterho@pacbell.net]:
Sure. Actually, you may have the legal right to do so. I don't know that it's ever been decided in case law that email belongs to the recipient, but I believe that's the case with smail. Balanced against that, of course, is the fact that the creator of a document automatically has a copyright in that material. I think that as webmasters we just have to use common sense. If
someone mentions in a message that they do not want that message published
or that they want to remain anonymous, I honor that request. Also, for
stuff that was clearly intended for me personally rather than me as
webmaster of www.ttgnet.com, I don't publish without asking permission
unless the message is clearly innocuous and would not embarrass the sender
if posted. * * * * * This from Paul Robichaux [paul@robichaux.net]:
Thanks. I'll check it out when I get a spare moment.
|
|
|
Wednesday, 1 September 1999 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] I'm starting to feel a bit under the gun. I'd originally planned to reach 100% completion on this hardware book February 28th, and here it is September 1 with me still working away on it. As always, O'Reilly is much more concerned about quality than deadlines, but even so I need to get this book finished. One thing I've decided to do is change how I update this web page a bit. In the past, I've updated it each morning and then gone to work on the book. From now on, I'm going to go to work on the book first thing, and update this page in the evening. At this point, I don't know when in the evening. It could be sometime after dinner (say, 7:30 p.m. Eastern Time) or not until late in the evening. Either way, I'm going to try this for a couple of weeks at least. We'll see how much it helps my productivity. If it doesn't make any noticeable improvement, I may go back to updating in the morning. We'll see. Next update will appear here sometime this evening, and will have a lot of good mail in it. * * * * * Evening: Well, I did get quite a bit written today, but today was a "writing day" rather than an "experiment, test, and verify day". I've had a great deal of back-and-forth with my editor at O'Reilly about "useful information" versus "interesting information". Nutshell books require the former, and that is a lot harder to come by than the latter. Not to knock a future competitor, but here's how I see it. I looked at the most recent edition of Scott Mueller's Upgrading and Repairing PCs the other day. This is one of those 1,400 page monsters, and it looks comprehensive at first glance. It's generally regarded as the best of the competing PC hardware books. And yet... I read his material on optical storage after I'd finished writing my own chapter on that topic. He goes on for pages and pages about things like the history of optical storage, how CD-R discs are manufactured, and so on. All interesting stuff, but not very useful when it comes down to brass tacks. What he doesn't do is cover the stuff you really need to know if you're buying, installing or using an optical drive. So, although the book looks to be comprehensive at first glance, when you read it you find that there's really not much there that you can put to practical use. And Mueller is by no means alone. When you start reading these books in detail, you find that all of them (or at least all I've seen) are pretty short on the nuts and bolts stuff. Having just written a nuts and bolts chapter, I know why. To write my chapter, I actually ended up building what amounted to four separate systems--one each running a SCSI CD-R under Windows NT, a SCSI CD-R under Windows 98, an ATAPI CD-R under Windows NT, and an ATAPI CD-R under Windows 98. Maybe it's just me, but the only way I know to find something out for sure is to do it. I've begun to wonder if some of the folks who've written these huge PC hardware books have ever really done the stuff that they're talking about. The flip side of all this, of course, is that it takes time to actually do the stuff. A lot of time. When I was writing software books, I was accustomed to knocking out a rough draft chapter in a week to a week and a half. With hardware, I'm running more like 3 or even 4 weeks per chapter. At least my editor knows me well enough to know I'm not sitting on my hands. But the book will be better for what I'm doing, and I hope the sales figures will reflect that. * * * * * Here, in a case of man-bites-dog, is something you won't see often. A picture of a vicious sheep herding a whole flock of Border Collies into a pen... * * * * * This from bdenman [bdenman@ftc-i.net]:
Yes, thanks. I already did that, and there is no file association listed for .png files. Anyway, the problem is not file associations. What I want IE to do is display the PNGs natively, which it does on all of my machines except my main workstation. * * * * * This from Tom Syroid [tsyroid@home.com]:
Oops. Sorry. Paul mentioned the site only in the subject, and I forgot to post that. It's www.atomz.com, and I haven't had a chance to look at it yet. Incidentally, rather than using a link in your sig, perhaps you should just use the actual URL. The link doesn't come across too well when someone pastes or replies. * * * * * This from Louis Hamos [lione@jps.net]:
Thanks for the kind words. As far as my own billing rates, I don't have them posted as Barbara does. I bill straightforward stuff (setting up a simple network, etc.) at $150/hour and more complex stuff (internetworking, etc.) at $250/hour. * * * * * This from Werth, Timothy P [timothy.werth@eds.com]:
I'm sorry. I wasn't clear. The .png file association is turned off in File Types. What I can't make happen is getting IE to display the .png file natively rather than trying to download it and invoke a viewer. It appears that having IrfanView set as the default .png viewer somehow "poisoned" IE for ever displaying .png files natively again. I know this isn't a problem with IE5 in general, as it displays .png files fine on other systems around here. * * * * * This from Jim Griebel [jgri@earthlink.net]:
Well, as far as I know I don't have Quicktime installed, so unless something else installed it for me without asking, most of my systems are displaying PNGs natively in IE5. Of course, it may be that my systems are faster than yours, and that Quicktime splash screen is disappearing too fast to be seen. As far as the thisweek.html link, I can't see any problem either on my local copy or my web server. Unless someone else reports the same thing, I'll assume that it's something odd at your end doing it. * * * * * This from Bill Anderson [andersonbill@earthlink.net]:
Thanks for the kind words. I just forgot to add that link, but it's fixed now and will show up the next time I publish. * * * * * This from bdenman [bdenman@FTC-I.NET]:
Well, apparently IE5 has native support for PNGs, but I sure can't make this version on my workstation display them without a helper application. * * * * * This followup from bdenman [bdenman@FTC-I.NET]:
Yes. I hate it when I feel as though I don't know what my software is doing. I really hate it when I can't figure out how to make my software do what I want it to do. And both of those situations increasingly obtain with most of the Microsoft software I use. * * * * * This from Frank A. Love [falove@home.com]:
Hmm. Apparently older versions of IE and Navigator required a Quicktime plug-in to display PNGs. That's apparently not the case with the versions I'm using--IE5 and Navigator 4.05. Why IE5 used to display PNGs natively and won't do so now, I have no idea.
|
|
|
Thursday, 2 September 1999 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] How do I hate Microsoft? Let me count the ways. I hate them to the depth and breadth and height my soul can reach... Last night, I was updating this page with FrontPage 2000. I made the changes, saved them, and exited FrontPage 2000. A few minutes later, I double-clicked the FrontPage 2000 icon as usual to fire up FrontPage and publish. FrontPage 2000 came up as usual, but immediately displayed the Windows Installer dialog. This dialog had my name already filled in, but not my initials or the serial number. I clicked Cancel and FP2K terminated. I fired it up again. Same thing. That was the straw that broke the camel's back. I will confess that I thought Pournelle was being a bit hard on Office 2000 with his recent stinging indictments of it. I no longer feel that way. I sent him the following:
I told Barbara that that was that. I've had it with Microsoft. I wish I could get rid of all their operating systems and applications immediately, but that's not practical, not least because I have two Windows 2000 books under contract. But I decided to do what I could now. That meant removing Office 2000, which is a Herculean exercise in itself. I finally got all vestiges of Office 2000 removed. That left me without mail, so I installed the latest version of Pegasus Mail, which I figured would hold me at least until I figured out what to do permanently. The problem, of course, is that all my stored mail is in Outlook format. Microsoft products are like the Hotel California. They make it easy to get it, but once there you'll never leave. I decided that I'd use the brute force approach: fire up Outlook on one of my other machines, call up my PST file, and individually forward every single one of my old emails--more than 50,000 of them--to myself. I'd then download them with a good mail package (like Pegasus) and run a little utility called FIXHEAD.EXE that strips out the forwarding headers and leaves only the original headers. But I digress. Before I did all that, I needed to have a word processor and spreadsheet that would read Office 2000 files. So I installed the latest version of Corel Office Suite. According to Microsoft, they made no file format changes between Office 97 and Office 2000. They lie. Once I got Corel installed, I tried opening a Word 2000 document in WordPerfect 8, which recognized the file as a Word 97 document. WordPerfect immediately GPFd. I exited and restarted Corel WordPerfect and tried opening several other Word 2000 documents. This time, it opened them just fine. The only problem was that the documents were empty. Just to verify this, I tried opening more than a dozen Word 2000 documents. Every one came up fine, and every one was empty. I then tried opening a dozen Word 97 documents. Every one opened fine and I was able to view, edit, and save the contents normally. Excel doesn't have this problem. Corel QuattroPro opens both Excel 97 and Excel 2000 spreadsheets normally. Well, if I won't use Office 2000 and I can't use Corel Office Suite, that leaves me with reverting to Office 97, which at least more-or-less worked. So I dug out the Office 97 CD and ran the install for Word 97 and Excel 97. That appeared to proceed normally, as did the install for FrontPage 98. Keeping my fingers crossed, I fired up FP98 and loaded the local copy of my web site, which I still hadn't been able to publish. I told FP98 to publish, pointed it at my pair web server and hoped for the best. As it turns out, it did publish, although it insisted on publishing all 300 files as though they were all changed files. Oh, well. I can live with that. I then attempted to install Outlook 98 (I need *something* to read the PST files) from the installation folder on my hard drive. I'd installed it from there numerous times, but this time it refused to install, saying that it couldn't find the distribution files it needed on www.microsoft.com. Huh? In the past, I'd successfully installed Outlook 98 from this folder even when no system on my network was connected to the Internet. Why not this time? I hate Microsoft. The only alternative seemed to be to re-install Outlook 2000, which I did with my fingers crossed. I'm hoping that that one Office 2000 product will live happily on my system without screwing anything else up. This all left me with the original versions of Word 97 and Excel 97, of course, so I had to go download the SR1 and SR2 patches to Office 97. As usual, Microsoft makes that an obnoxious task, by demanding that you accept cookies and forcing you to provide personal information. That's unacceptable. At any rate, I let SR1 and SR2 download overnight, and installed them this morning. Why they didn't make SR2 include SR1 updates, I'll never know. The installs seemed to go fine, and I now have functioning versions of Word 97 and Excel 97--both of which open the Office 2K versions of documents just fine, thank god--FrontPage 98, and Outlook 2000. And all it took me was about 8 hours worth of work to fix a problem that Microsoft caused. I hate Microsoft. If you haven't installed Office 2000 yet, don't even think about doing it. This is the worst piece of garbage that's ever been foisted on computer users as an allegedly finished product. It's not. I don't even think it qualifies as a beta as I understand that term. More a late alpha. Microsoft delenda est. * * * * * Pournelle posted a couple of my messages from yesterday in Mail today, along with a detailed response. I sent him the following in response to his response:
We certainly agree that Office 2000 is not-ready-for-primetime, but I think it's more than that. I think the problem is that Microsoft's reach has finally exceeded its grasp. They have finally succeeded in making something so complex that it may well prove impossible to debug. Actually, Office 2000 isn't even the best (or worst) example of this. Windows 2000 is still an incredible mess. The Professional version looks to me like an early beta at this point (not an RC in the original sense of the word), but the real problem is with Windows 2000 Server, which still has huge chunks that are not working correctly or are missing entirely. I know that you think highly of Windows 2000 Professional, but just try, for example, to use it as a DHCP client on an NT4 DHCP Server. You may find, as I have, that Windows 2000 Professional grabs an IP address that is already being used by another system on the network. With Windows 2000, Microsoft has redefined DCHP from Dynamic Host Configuration Protocol to Dynamic Host Conflict Protocol. Windows 2000 Professional has done this to me both times I've installed it, on two different machines and on two different networks. I'm not prepared to say that the problem is reproducible (few behaviors, good or bad, are reproducible with Microsoft software), but it seems to me that this is a problem that should have been caught and fixed several builds back. It's been said that Microsoft is betting the company on Windows 2000. If so, there must be quite a few people at Microsoft right now that are looking for a way to hedge their bets. I'm sure that if they could buy NetWare or Linux, they'd just do so and have done with it. The irony is that I have two Windows 2000 books under contract with O'Reilly. I think I'll start the introduction, "If you find yourself sentenced to administer a Windows 2000 Server network, please accept our sincere condolences. Assuming that you can't somehow convince your superiors to use a real network operating system like NetWare or Linux instead, here's what you'll need to know..."
Again, I agree. But this is not the first release of Office. Depending on how you count, it's at least the fourth. By now, they should have their code fully debugged. Instead, they have a complete mess. If Corel Office Suite 8 could open Word 2000 documents, I'd have converted to it by now. I'm also seriously considering downloading StarOffice (64 MB) to see what they have to offer. Of course, Microsoft takes the Hotel California approach to designing software--make it easy to import stuff from other software; make it impossible to export from Microsoft applications. It may just be worth biting the bullet and getting away from Microsoft entirely. Microsoft delenda est. * * * * * This from Robert Rudzki [rasterho@pacbell.net]:
I discounted the likelihood of direct inductive or RF interference based on the fact that two different PCs were behaving similarly in different areas of the home. My guess is that he'll find that he coincidentally had both PCs connected to the same circuit and that that circuit was sagging. If that's the case, simply connecting to a different circuit should solve the problem. If for some reason he has to use the existing circuit, adding a line-interactive UPS will at least eliminate the symptoms if not the problem itself. * * * * * And this, also from Robert Rudzki [rasterho@pacbell.net]:
Very interesting. Brings new meaning to the word "microcomputer." * * * * * This from Brian Bilbrey [bilbrey@pacbell.net]:
I can certainly believe that. The problem, of course, is that I have only the existing GIF images in most cases. So, the question is not whether to create the images afresh in PNG format or convert existing GIFs; the question is whether to continue to use GIFs or to replace them with PNGs that have been converted from the GIFs. * * * * * This from Bo Leuf [bo@leuf.com]:
Well, perhaps the problem, as I noted yesterday, is that my machine is fast enough that I can't see the splash screen. But it sure appears to be displaying PNGs natively in IE5. I'll probably look at Opera sometime again, but I wasn't impressed the last time I looked. It did little or nothing for me that IE5 doesn't do, was slower than IE5, and had far fewer useful features. There are a few things I liked about it, but on balance I'll stick with IE5 for now, much as I hate Microsoft. * * * * * This from bdenman [bdenman@FTC-I.NET]:
Thanks. I'll look at it when I have a moment... * * * * * This from Tom Syroid [tsyroid@home.com]. I did ask permission before posting it here...
Exactly. * * * * * This from Jerry Pournelle [jerryp@jerrypournelle.com]:
Not me, boss. I advised you to change over from Word to FrontPage on the grounds that Word would rapidly run out of steam for maintaining your web site, and that FrontPage was the easiest thing I knew of for someone who wanted to just get a web site up and running without worrying about HTML and other details. I still feel that way. I'd never count on Microsoft to integrate anything with anything. I think the fundamental problem is that their roots are in standalone PCs. When they build something, they're still building for individual PCs rather than seeing the network as the computer. Witness their pathetic flat namespace domain schema in NT4 (versus NetWare's hierarchical directory service, which they implemented years ago with NetWare 4.0). Speaking of which, here's a deep, dark secret that Microsoft doesn't want anyone to know. Windows 2000 *still* uses domains. All Active Directory really is is a "directory service view" of an underlying flat domain model. AD as implemented in Windows 2000 is less capable than NDS was in NetWare 4.0. In fact, I suspect many purists would agree with me that AD doesn't deserve to be called a directory service at all. Microsoft delenda est.
|
|
|
Friday, 3 September 1999 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] What a day. Here it is after 10:00 p.m. and I'm just now getting around to updating my page. There's lots of mail, and I always try to give that priority. Barbara brought home a 10-week old Border Collie pup (picture here) that's been running us off our feet. More tomorrow, if I survive the night. I hope I can publish this successfully this time. The last two evenings, FrontPage 98 has insisted on publishing the entire web site rather than just changed pages. We'll see. * * * * * The following two messages both concern the material posted here that describes the big to-do about the NSA and back doors into MS encryption products. * * * * * This from Paul Robichaux [paul@robichaux.net]:
And this from Jeff Martin [jemartin@newsguy.com]:
Yes. If true, this is in fact very disturbing. I note that Pournelle says Microsoft denies that there are backdoors in their security, and he believes them. But I'm not so sure. Oh, I believe that Microsoft *thinks* there aren't any backdoors, but compared to the NSA, Microsoft is a big, friendly puppydog. It wouldn't surprise me a bit to find that NSA had programmers operating undercover within Microsoft, and it wouldn't surprise me a bit to find that backdoors do in fact exist. * * * * * This from Brian Bilbrey [bilbrey@pacbell.net]:
Thanks. I was aware that licensed applications produced licensed GIFs, but I think Unisys is playing on people's fears (or wish they could.) I'd be very surprised if they got many people coughing up their license fee. * * * * * This from [root@mail.bcsupernet.com] On Behalf Of cc:
Believe me, I'd switch to Linux if I had time to learn it, but I don't. That and the fact that I have two Windows 2000 books under contract with O'Reilly. I suppose it's no bad thing to have a book written by someone who's not a fan of the software. * * * * * This from bdenman [bdenman@FTC-I.NET]:
More than a tad, thanks. I just installed StarOffice and we'll see what transpires.. * * * * * This from Jim Griebel [jgri@earthlink.net]:
Yep. Your symptoms seem exactly like mine, so your solution probably would have cured my problem. Of course, I now have Outlook 2000 re-installed, for as long as it continues to work. I read of a similar situation over on Pournelle's page today. A reader had installed Office 97, installed SR1, and then attempted to install SR2. Setup refuses to install SR2, claiming that SR1 had not been installed. I had exactly the same thing happen to me some months ago, and I suspect a DLL conflict was the cause there, too. * * * * * This from Robert Rudzki [rasterho@pacbell.net]:
Sure. Feel free to point your students here, although I certainly don't hold this up as any kind of worthy example. I just keep things as simple as I can. * * * * * This from J.H. Ricketson [culam@dnai.com]:
Thanks for the kinds words. The nav bar at the bottom of Sunday is a good idea, and I've added it. As far as the Celeron uniprocessor thing, the latest I've heard is that Intel still hasn't actually done anything about it, but simply announced that they might. They may well disable SMP support on the forthcoming 100 MHz FSB Celerons, but for now they seem satisfied simply to label the package "Uniprocessor Only". As far as Clarke, I rather liked Gregory Benford's Corollary to Clarke's Third Law: “Any Technology Distinguishable From Magic Is Insufficiently Advanced.” * * * * * This from Dave Farquhar [farquhar@lcms.org]:
Thanks. Actually, I've uninstalled Corel Office for now. I downloaded StarOffice today (all 65 MB of it) and installed it a little while ago. It's definitely different, but it does open Word 2000 documents natively. I'll definitely keep your suggestions in mind, though, because I'll probably reinstall Corel Office at some point. * * * * * This from J.H. Ricketson [culam@neteze.com]:
Thanks. I'll definitely keep that in mind. For now, though, I'm so covered up that I don't have much time to mess with stuff that's not absolutely necessary to what I need to do. I mostly just keep images turned off when I'm browsing nowadays, and that cures most of the problems. * * * * * This from Paul S R Chisholm [psrc@lucent.com]:
Thanks. This time, at least, I was able to get Office 97, SR1 and SR2a installed successfully. I think I'll let sleeping dogs lie for now, because I hope that Office won't be on my systems much longer. If StarOffice works out, that could happen sooner than I think. * * * * * This from J.H. Ricketson [culam@neteze.com]:
I've come to like SCSI a lot better recently than I did for a long time. In fairness to ATA, you can in fact do a low-level (or pseudo-low-level) format on any recent ATA drive. Every manufacturer I know of supplies a utility that writes all binary zeroes to their ATA drives, which addresses the main reason for doing ll formats these days--wiping out viruses, deleting damaged partitions, etc.
|
|
|
Saturday, 4 September 1999 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] Well, FrontPage 98 insisted on publishing the entire web site for the third time in a row. Barbara asked me last night what would happen if I did the edits under FP98 on my system, and then published with FP2K from her system. Good question. I'm going to find out as soon as I finish writing this. Nope, that didn't work. I couldn't get the web to publish at all from Barbara's machine under FP2K. Apparently, something has convinced FP that my local web is no longer a web or something. I'll try again from my machine, and hope it doesn't insist on publishing the whole thing yet one more time... * * * * * And, with PNGs on the brain, I managed to save the picture of Gypsie, the new Border Collie pup as a PNG rather than as a JPG. Big mistake. The PNG was 171 KB. I've just converted it to a JPG, which is 50 KB. * * * * * This from tek1@inwave.com:
Well, I already did comment on it in yesterday's journal, although that hadn't been published yet when you sent this message. Basically, I have a sneaking suspicion that there are in fact back doors in Microsoft's (and others') encryption, that the NSA put them there, and that the software vendors aren't even necessarily aware that they're there. * * * * * This from J.H. Ricketson [culam@neteze.com]:
Yes, there was a long hiatus in my experience. The last time I worked with SCSI, it was a mess. Adaptec had already become the dominant force in the industry, but there were several strong competitors. ASPI was in its early days, and SCAM didn't exist. Connecting SCSI devices was an adventure, with incompatibilities between different manufacturers' devices (that supposedly used the same SCSI standard), conflicting drivers and firmware revs, and operating systems that had very poor SCSI support, if any. In the meantime, ATA just kept getting better, faster, and cheaper. I saw no reason to use SCSI for anything other than servers and high-end workstations (in the true sense of the word). Nowadays, that's all changed. The standards are solid, SCAM makes SCSI truly Plug-'N-Play, and the price premium that SCSI has always maintained over ATA is dropping. SCSI is still an incremental expense, both because you have to buy a host adapter and because SCSI devices are still somewhat more costly than ATA, but it's now a reasonable alternative for performance-oriented systems. For many years, I had no SCSI devices in the house. Now, I have three systems running SCSI host adapters, with a variety of devices installed, including a scanner, hard disks, and CD-ROM and CD-R(W) drives. I'll soon have a SCSI DVD-ROM drive and a SCSI DVD-RAM drive running on another system I'm building. I still maintain that ATA is the best bet for ordinary users, but for the kind of people who read this page, SCSI is something worth having. That's a 180 for me. Until very recently, I was of the opinion that SCSI made sense only for servers. I'm glad my readers convinced me to take another look at SCSI. * * * * * This from Paul Robichaux [paul@robichaux.net] regarding Windows 2000 Build 2114:
Thanks. I was kind of expect to get RC2 sometime next week, and was planning a system around it. I guess that means I can put off building that system for another week or two. * * * * * Late Morning: Now here's something truly frightening. CNET reports that Russian software used to run their strategic warning system is not Y2K compliant because they ran out of money. So, come 12/31/99, they'll have about 2,500 thermonuclear warheads pointed at us and warning software that hasn't even been tested, let alone debugged. We've lived with the possibility of being nuked intentionally for forty years now, and that's terrifying enough. Somehow, the possibility of being nuked unintentionally is even worse to my mind. "Oh! Sorry, our mistake..." The US has given Russia billions of dollars. I've said all along that instead of giving the money to them with no strings attached, we should trade them our money for their warheads. The problem, of course, is that Russia is barely a country at all nowadays. By most measures, it doesn't even qualify as a third-world country. Fourth-world, perhaps. And with all of that, they cling with pride to the nuclear trinkets that once made them a superpower. Now that Russia is essentially ruled by various Mafya gangs and the government can no longer provide even such basics as potable water to the residents of their major cities, something needs to be done to remove these dangerous toys from their control. I used to think that the Y2K fanatics who quit their jobs, sold their homes, and headed for the hills were over-reacting. Now I'm beginning to wonder. 2,500 nukes means that there are probably fifteen or twenty targetted on the Winston-Salem/Greensboro/High Point metro area, and all it takes is one. I can see it now. As we watch the ball drop in New York at midnight, the TV will die, the lights will go off, and then it'll suddenly get very bright outside. I'm going to be seriously annoyed if that happens. I hope the Defense Secretary talks Russia into removing the distributor caps from their nukes.
|
|
|
Sunday, 5 September 1999 [Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] It's a beautiful, gray, rainy morning here. The remnants of Tropical Storm Dennis came started coming through last night. We've had about 3 inches (7.5 cm) of rain so far, and it looks as though we may get at least an inch or two more. This area has been suffering from drought, so this rain is very welcome. I hope the people on the coast didn't get nailed too badly, though. * * * * * Some months ago, I declared Rambus memory moribund, and it looks like that prediction is coming true. The Register has been running numerous stories lately about the rush away from Rambus, including this one. * * * * * This from Fred Mora, sent to Jerry Pournelle and CC'd to me:
I'm at the point where I would love to abandon Microsoft operating systems and applications. Realistically, I can't do that. I simply don't have time to learn all new tools. However, I am gradually getting away from Microsoft to whatever extent possible. I've installed (and subsequently uninstalled) Corel Office Suite, and am now looking at StarOffice. That seems like a product with real possibilities, although I am concerned about Sun's announced plans to convert the whole thing to a thin-client environment. I prefer it just as it is, as I suspect many people will. The whole thin-client thing is really just Sun's attempt to do in Windows. Better they should focus on Linux and give up the thin-client garbage. * * * * * This from Francisco García Maceda [maceda@pobox.com]:
Well, although I agree that hard drives are the one SCSI market segment where the SCSI price premium remains most noticeable, I think you're comparing apples and oranges to some extent. For example, I find the Seagate Medalist Pro 9.1GB UDMA (ST39140A) for just under $200. That's a 7,200 ATA-33 drive, with 9.5 ms average access and 512 KB cache. Compare that to the Seagate Barracuda 9LP 9.1GB (ST39173W) for just under $300. That's an ultra wide SCSI drive that also runs at 7,200 RPM, but has 7.1 ms average access time and double the cache. The SCSI drive costs about $100 more, but that's for a faster drive that runs ultra wide SCSI versus ATA-33 and has twice as much cache. That drop from 9.5 to 7.1 ms is a pretty expensive one to make. I don't know offhand of any ATA drives with 7.1 ms access, but if they exist, I'll bet they'll be noticeably more expensive than a 9.5 ms version of the same drive. So, how much of the $100 price difference is due to the faster mechanism, how much to the doubled cache size, and how much to the SCSI interface? On stuff other than hard drives, the premium is even smaller. For example, I see that current Toshiba ATAPI CD-ROM drives go for about $45, where a Plextor SCSI CD-ROM drive goes for a $25 or $30 premium. The Plextor is nominally slower, at 32X, but in practice is considerably faster than the ATAPI drives, particularly for stuff like Digital Audio Extraction, where speed really matters. For example, I have a 36X Toshiba here that rips at about 6.8X. The 32X Plextor rips more than twice as fast. I remember some years ago when the SCSI premium was huge. Exactly equivalent drives sold in ATA for $500 and SCSI for $1,500. So, I don't disagree with you that SCSI sells at a premium over ATA. My original point was that that premium seems to be dropping, and I continue to think that's true. * * * * * This from Bo Leuf [bo@leuf.net]:
I don't pretend to know much about the space program, but I think Pournelle's continuing indictments of NASA have the ring of truth. This is just one more nail.
[Last Week] [Monday] [Tuesday] [Wednesday] [Thursday] [Friday] [Saturday] [Sunday] [Next Week] |