Standards: It's amazing how many people send around Word, PowerPoint, and Excel documents. Why do they expect me to be able to read these? Do they expect me to cough up ~ $500 like they did (or didn't) for software I don't need other than to accomodate their whimsy? If they send me an OpenOffice document instead, there's no financial burden on me and it's an open standard which I am in some sense more entitled to use/interpret/read.
As for Firefox -- since this is now popular enough at my workplace, I basically don't need to test for compatability with IE (which is difficult to run in Linux), and I don't support it in my web projects. If it works in Firefox, Opera, and Konqueror (which almost tests Safari compatability), then it's good enough for me. I am also looking forward to more complete support for things like MathML, which will gradually make life much easier for me. As for IE support, I wish there was a web page that launches an ActiveX script that installs FireFox with little notification.
For an open-source project, the number of developers tends to increase with the number of users. I don't think it's a linear relationship, but it's certainly monotonic. And when talking about open-source, the distinction between developers and users gets wonderfully blurry.
Finally, I love the wide variety of open-source projects going on! I often find projects that are useful to me, and it seems like each year computing just becomes easier thanks to OSS. Better programming languages, more libraries, more complete hardware support, improved documentation, etc. The more people become aware of open source, the more they will get involved in it. This is of direct benefit to me, and everyone else too!
I think your underlying assumption is that if all drivers act in their own best interest, the total efficiency of the system maximizes (I mean that the average driver's trip time minimizes).
It's not obvious to me that this assumption is true. It may well be the case that the best overall traffic pattern involves sub-optimal routes for some drivers.
I think FreeCiv is a leading example of how OSS can enable collaborative game balancing and development. For a moment, don't think about the software at all. Just look at the game mechanics, and how FreeCiv has been fine-tuned by a community of interested players. That's really kind of new! The typical game development model these days is that the game company works on balancing totally in-house, and might to a limited extent actually listen to the players. The result is often frustrating to fans of the game who would like to try out their own ideas.
FreeCiv definately leads the pack with a large active community and a long history of development. I don't think you can count nethack in this class because, AFAIK, its core development team has remained an exclusive group.
Yeah, that quote from Roper hit a nerve with me too.
It's too bad that the term "RPG" now has this dual meaning. The two definitions are worlds apart. It means either "computerized stats-based avatar-building game", or "role-playing game". The former refers to strategy games like Diablo where players try to perfect chcaracter builds and collect ever-more fantastically rare and powerful items to either compete against each other or to make the game easier and easier.
That, obviously, has nothing whatsoever to do with role-playing where the objectives are story telling, creating an atmosphere, and developing a believable personality. I'll argue that Neverwinter Nights was the first computer game which aspired to be an RPG, as it was the first game to really allow for a DM.
On the pro-Linux side, I had a night-and-day experience upgrading one of my computers to a new MB/CPU/video card. It dual-booted to Windows 98 and Linux. I wanted to keep Win98 because it's a computer that my young daughter uses to play a bunch of kids' games. Upon first booting up, I went into Linux. Perfection! Zero work required. I was upgrading from an really old (RIVA TNT2) NVidia card to an old NVidia (GeForce3 Ti) card, which made life easier for Linux.
Then I tried booting into Windows. And tried, and tried. For a long time I had to turn off "plug-and-play OS" in the BIOS because Windows kept all the incorrect BIOS drivers. I even had to fight to get the CD-ROM working so that I could give it the Windows install CD to loads dozens of drivers off of. Every single piece of hardware required 1-4 reboots to get working, independantly. Honestly I struggled with this for about a week, a couple hours a day. Among other things I had to re-assign the CD-ROM drive letter manually three times during the week, each time so that it could find the $@@#%&* Windows CD. Unbelievable! Admitedly Windows 98 is old and barely supported these days, so to be fair I should compare it to, say, RedHat 5.2 or something like that. But even Windows XP is not a piece of cake when dealing with massive changes to hardware.
Okay, my other comment is that the newbie's experience with Linux (and I mean Linux newbie, not computer newbie) varies a whole lot depending on what hardware they are using, and what they are wanting it to do.
For example, I helped a friend install Gentoo Linux on an Athlon 64 laptop, going for the full 64-bit OS and applications. It works now, but it was not easy. It took a while to get the broadcom wireless card working with ndiswrapper, get accellerated X with his ATI video card, etc.
You can't really blame Linux developers for difficulties supporting the latest hardware. Microsoft doesn't have nearly as hard a time since they are so popular that all hardware manufacturers write their own Windows drivers. Not only do Linux users have to write their own drivers, they often can't get the specs and have to reverse engineer it. It's a major testament to the sheer power of the open-source model that so much hardware is supported in Linux, often with better reliability than the Windows drivers!
Then there's a lot of powerful Linux software with a steep learning curve. MythTV is kinda hard to set up right. Doing firewalls well with ipchains requires a lot of (worthwhile) reading. Working with video using transcode is also time-consuming and involves lots of trial and error.
So for the typical office computer user using well-established hardware and a nice easy distro like Mandrake or Knoppix, Linux can be substantially easier to use and learn than Windows. But the power-user or hobbyist will find no end to the cool things that Linux can do with a lot of hard work. And someone running Linux should not expect to buy a GPS device + MS Streets software and literally plug and play!
While we're bashing Mac mice, I still vividly recall the negitive impression the circular iMac mouse had on me back in '97 or so. One of the campus public computer labs had some nice scanners that I would use every few months, and they had just updated their Macs to the colorful, translucent iMacs. There was no indentation for my fingers on the one button (I too feel like I'm wearing a mitten on a Mac) and the mouse was perfecly axialy symmetric, and so there was no way to feel the orientation of the mouse in my hand. Every single time I went for the mouse to click on something, the cursor would go off in some random direction and it would take me a few seconds to re-orient myself!
I always hear that Apple holds aloft the banner of greater usability and productivity, but I just can't take seriously any company that would let a mistake like that out the door. Had no one even tested the mouse for three minutes before mass-producing it?
Okay, that was a long time ago. But it was so disconcerting that I still think of it whenever I hear about Mac mice.
From what I hear, the argument for the one-button mouse is that some small fraction of new users will experience confusion and frustration with something more complicated. But surely that confusion would not last too long? So Apple really wants the majority of their users to buy another mouse to replace their functionally-challenged one? What about public computer facilities where the admins generally go with the original equipment? I'm glad they may start making two-button mice, but that scroll wheel is very handy, too. I don't personally like the side buttons because my thumb and pinky are used to grip the mouse and I find that I click those buttons by accident a lot. Also, the extra functionality of the side buttons is very limited in scope, and can be conveniently replaced with things like mouse gestures.
I agree, www.silentpcreview.com does have a lot of good advice. But like any review site, they only cover a small fraction of the current hardware available that is marketed to the quiet PC crowd.
I spent some time and money last month going through that site and working to quiet my PC down. I set up a new PC to be always-on in a corner of my dining room. The location was dictated by cable access and wi-fi reception. (It's running Gentoo Linux and is serving as a gateway/firewall, records radio shows, web server, etc.) The dining room has hard floors, and the PC noise was absolutely unbearable at first. Now it's just about as loud as the 700W APC UPS it's on, which is noticeable in that room but not at all obnoxious.
* case: Evercase GC4292: inexpensive steel case with no P/S and can handle a 120mm case fan. * P/S: 300W Seasonic supersilencer -- has a very quiet fan, I really like it. * CPU fan: Arctic Cooling Copper Silent 2 TC -- I'm very happy with this too, barely audible with case open. * VGA fan: Zalman VF700-ALCU Ultra quiet fan -- Another winner, though the RAM heatsinks were unnecessary (I'm not overclocking) and didn't even all fit without compromising the main heatsink's contact with the chipset. * I mounted the HDD with rubber grommets. The case's HDD bay supports these but I had a hard time finding the kit to go with it. I'm still looking for more grommets. I believe that this helped substantially, but I don't have test equipment and I swapped hard drives at the same time. The old drive without grommets was a prominent noise source, and now I don't hear it except when the head is seeking. I don't remember the brand/model, but I think it's a Maxtor 5200RPM, 200G drive. * I am still struggling with a quiet case fan. This is now the only component I can really hear with the case closed up. I tried two Antec variable-speed fans with thermistors on them, one 90mm and the other 120mm. I'm currently using the 90mm fan because the 120mm fan has a louder and more annoying pitched hum. I like the fan speed control idea in this article.
I also bought the Antec noisekiller kit in a whim, but it was difficult to use and I don't think it really helped any. I was hoping for drive mounting screws and grommets, but they were not included.
While there is a lot about the US tax laws that I find ridiculous, I at least have some appreciation for the difficult problem of fair taxation. We have hundreds of millions of people trying to save on their taxes, and if any loopholes exist, they will be found and exploited. While I would love a simpler system, there is balance. If it's too simple then it's easy to abuse, and if it's too complex the same could be true. An abusable system is worse because you'll either pay more than your share, or you'll pay a tax preparer/crook to find sneaky ways to save money. You want a system that doesn't demand a paid tax preparer, and thus you want a more airtight one.
We are already in this bad situation somewhat, but I think that while it could easily be worse, it's not easy to make it much better than it is now.
I also don't think that the tax preperation lobbyists/donors have more influence than the tax-payers who want an easier system. We outnumber them. So the argument that the tax-preparers force the IRS to keep them in business doesn't make sense.
One of my pet peeves is the tax bracketing system. One can save a fistful by managing to squeeze into a lower bracket. This encourages abuse. To make taxation a continuous, monatonic function of earnings requires no more than grade-school mathematics. But it would require just slightly more math than the present system: multiplication followed by subtraction instead of straight multiplication. So the IRS deems that the subtraction pushes the requirements into the lair of advanced mathematics, and thus the idea doesn't make the simplicity cut.
It's our own fault, really. We ask for simpler tax law and this is what we get.
Re:One of my favorites is "Scotland Yard"
on
Fun Tabletop Games?
·
· Score: 1
I can second both of these games. Scotland Yard is such an original game concept! Lots of fun.
Empire builder and follow-up crayon rail games are also great. I think Eurorails is my favorite, since the games often come out unexpectedly close with the winner only a turn or two ahead of the rest even though their tracks and strategy are completely different from other players. I've always wanted to computerize these rail games but haven't found the time yet. The main reason would be to play with friends and family that are living 3000 miles away. And partly this would be fun just to collect statistics to refine track-building strategies and work out good distributions of demand cards. Also it would be a fun game to design an AI for.
I heard one study recently demonstrating that in grade school, girls do better in science and show more interest then boys. In high school, men edge out the women, in college (for the hard sciences anyway) it's more like 3 to 1 men-to-women, and in grad scool it's more like 10-to-1. One interpretation of this is that innate interest or capability is not the problem, but social pressures are. Is this a problem? Yes! Even if for no other reason than that science needs diversity of thought to make the most progress.
A non-PC, unpublished, didn't-officially-happen study I heard about was histogramming grade distributions for the colegate physics 100-level course series (a couple thousand students). The men's grades fell into a neat bell curve with maybe a B average. The women's grades were a remarkably flat, uniform distribution from 0 to 100 percent (where F was about 50 or 60) -- remarkably distinct shape from the men. I'm not sure what conclusion to draw from that, but in my oppinion it's a serious problem.
No, not everyone should excell at sciences, and it's certainly true that the hypothetical average woman has different strengths and weaknesses than the average man. But the 100-level physics courses should not require extraordinary ability to score well in (just hard work at worst), and I have no reason to expect that women are naturally worse than men at it either. I also think that someone's aptitude for sciences depends primarily on their own self-expectations.
My first reaction to this was that this MUST be a hoax. It's just the kind of thing my friends and I would have joked about in high school. For the record, we also came up with the McCow which is a rectangular (well cuboid) bovine that's stackable. These occasionally get loaded onto trolley cars and driven through an astroturf pasture for the visitors. But that's another story...
My second reaction is that for this to be in any way scientific, we need free access to the source code. It sounds like we're presented with basically one alrogithm that converts a song into 2 or 3 numbers, and these are compared with the distribution of other known-to-be-popular songs. Well, what happens when these algorithms change? How stable are the results against using different algorithms or different constants or different datasets of known-good music? The lack of this kind of in-depth analysis (which would have to be done by an independant party to have any validity) calls into question the whole technique.
I still don't like the idea of a computer algorithm helping to determine what future songs get airplay, but maybe it's not any worse then our present situation. Today we have "top-40" stations with play lists less than 20 songs deep, and big-wigs that listen for the hook in the first 15 (IIRC) seconds of a song.
Is there any open-source backup software for Windows? I have a little Perl script that I've been using for my laptop running Linux that uses tar for compressed, incremental, selective backups onto a public backup SMB share (or any mounted directory) after gpg encryption. I can "port" it to cygwin very easily and may use that if I don't find anything friendlier.
Oh, nice! I like the gallery program at first glance. I'll investigate it more soon.
I wrote my own pictures -> HTML gallery program a few years ago, using Perl and ImageMagick (via the Image::Magick Perl module). But it was never quite advanced enough to release it -- just a couple hundred lines of code (which accomplishes a lot in Perl). I still use it from time to time, often just to make thumbnail images for other purposes.
I usually bring my laptop, and show pictures on that. The screen is not large, but at 1600x1200 resolution the pictures look VERY nice! Admitedly not more than 3 or 4 people can really gather around a laptop, but it does the pictures justice. A can't imagine that a TV would be acceptable unless I only had a 1/4 MP camera.
I have to mention software, too. gqview. I haven't found any Windows software that even comes close to gqview. I suppose the best I've found so far for Windows is Irfanview, but compared to gqview it's still awful. Now I read somewhere that gqview works under cygwin. For my Windows box, I'll try that next.
Also, as for taking good pictures, I'm loving my Canon digital Rebel. I heard that professional photographers take around 100-1000 pictures for every one they like. I'm not a pro, but my standards are much lower. So I think my ratio is about the same.
Digital allows me to take pictures indescriminantly, and try all kinds of artistic shots that I wouldn't dare waste film on. I learn A LOT from these, and sometimes I get a gem. On a typical day hike I'll take a couple hundred pictures. Of those, I'll delete about a third (near-duplicates or badly flawed), really like a dozen, and might find 1 or 2 that are (for me) very good photos. I have a couple scripts that help me sort pictures into directories (using hard links), and these are ones I would show to people if I have the time.
I didn't like the article at all. I think there are several factual errors, trumping all the problems with the flash presentation.
Is a larger sensor really necessarily better than a smaller one? I don't think so, but it definately places different requirements on the optics. Perhaps megapixel-for-megapixel the larger sensors are sharper because of bleeding from one pixel to another? Anyway, it's the combination of optics and sensor that determines image quality (and I suppose post-processing, etc). DSLRs can have good-quality optics, but the larger sensor requires larger lenses, and that makes them more expensive for the same amount of quality.
Depth of field -- he claims that no camera can have the same depth of field as the human eye. Again, just not true. We are constantly focusing on things near and far, but at a given focus, we have shallow depth of field. An SLR camera gives you the most control over depth of field, and is generally much more narrow than a compact digital. With an artistic eye, this can be a great advantage of the DSLR.
I'll read the rest later. He does make some good points. But I was not impressed initially.
I hope that he mentions the shutter lag times and focusing time. Those aspects are very important for quick candid shots, and the DSLR's really beat out the compacts here.
All Blizzard's other games cost about $50 new, but there was no subscription fee and you could download a free demo. That sales model is fine with me. They talk about supporting the online community and adding more content, but they've done that with all their other games too, again w/o the subscription fees.
If they only charged the monthly fee, that's like a close-to-free ($15) limited-time demo, and if you like the game and play it for more than 4 months (Most Blizzard games maintain their appeal for years) then you've already spent more than $50 on it.
Also, I like online games that I can play with my wife. Again, most games don't really require me to buy two copies for LAN play. But now I'm looking at $100 PLUS $30 per month. That's nearly as much as a broadband connection (which also is usually free installation)!
I'm sorry. This game REALLY sounds awesome, and I would LOVE to play it. I'm totally into these kinds of games. But the cost is just way too high. That kind of monthly fee is much too large a fraction of my monthly savings.
Also keep in mind that Blizzard sold 250,000 copies on the first day. Suppose they've sold 1 million by now. That's 50M already, plus 15M per month. I think that's far more than enough to support the 80+ servers and staff. Blizzard is getting greedy.
First, after several tens of years, the composition of fuel rods changes significantly -- the shorter-lived components will decay and the waste will generate far less heat. The ideal storage environment changes substantially then.
Also our current waste-management techology is immature, and not proven to be good enough. But a few new developments are on the horizon.
Future technology is likely to make fuel reprocessing more economic (and I think he did this without even mentioning breeder reactors).
Finally, the Yucca mountain storage facility is gridlocked in politics and thus not a realistic short-term option.
Then the author addresses your question, and suggests using our present temporary solution of casques but upgrading it to a centralized facility that can be hardened against terrorist abuses. It's not clear to me that this is the best way to go but it's obvious that we need an *immediate* improvement over what we're doing now, and that we really want to consider a temporary storage technology that's good for 100-200 years, not necsessarily 100-200 thousand years.
I was thinking about this very recently. What most distinguishes journalism from science?
In a good scientific article, the author is often her own best critic. The paper will first establish the facts: precisely how the experiment was done, and what the results were. That's usually the easiest part. The great majority of time spent is in figuring out just how conclusively the data confirms the thesis. This involves looking for alternate explanations for the data, and guessing at any possible systematic bias.
After reading a good scientific article, you should come away with an understanding of what confidence there is in the conclusions. This study of uncertainty and systematics is really the most important part -- it's perhaps the defining characteristic of science! And this element is precisely what's lacking in journalism.
Every scientist will have a personal bias going into a study, but does their best to quench that bias with a strong and healthy skepticism. And a reader or journalist can, by reading multiple articles addressing a particular issue (independant confirmation is another cornerstone of science) gain an more precise and less-biased oppinion.
So, to be unbiased, a journalist doesn't have to be "ballanced" but they should compile as many differing oppinions as possible, and judge them by the strength of their evidence. Any report of science would be incomplete/broken/useless without these three elements: discussing the body of evidence, reporting the (statistical) uncertainities, and investigating for possible bias.
The articles refered to in the story do mention the strong personal biases of the fringe scientists. But it would be even more convincing to criticise their research, not their politics.
Journalists often have the hard job of distilling science for the masses. But I think they go way too far, presenting a collection of scientist's oppinions rather than presenting scientific evidence. If the focus were on the data and not the oppinions, the personal biases of the researchers and journalists would be largely removed.
To conclude, journalists absolutely must preserve: 1) The evidence -- most importantly what experiments were done. 2) Uncertainty -- how convincing is the evidence? 3) Bias -- how susceptable are the experiments to bias? 4) Independant confirmations -- not how many other scientists agree at a convention, but how many fundamentally-different experiments support the same hypothesis?
I was actually considering voting for Badnarik, on the premise that Instant Run-Off voting is actually a vitally important election reform for the US. The Republican party so far has the least incentive for establishing it. More votes for Badnarik would help with this. And I live in New York, so "my" electoral votes will go to Kerry either way.
But then I figured that the popular vote is also somewhat meaningful, and I should really only throw away my vote if I were relatively neutral between Kerry and Bush. I don't like either one really, but I think that Kerry would make a far less-awful president than Bush has been.
As for Kerry, although his anti-Bush statements have been on target and valid, his own oppinions on what to do in Iraq are unrealistic at best, pandering at worst. He might sound good, but I feel that many of his arguments have no substance. I'm also disturbed by Kerry's incessant trumpeting of Cheney's daughter. I don't know of any "real people" who would make such a big deal out of such a trivial fact, going out of their way at every turn to make polite-sounding comments that are thinly-disguised attacks like these. It just disgusts me. No decent human being would do that. Kerry did. Sigh.
But all of these issues that might in a normal elecetion swing my vote are suddenly made trivial by the outrageous actions of Bush. His war on Iraq was based largely on evidence of nuclear weapons, which we later find out was such weak evidence that it was more nearly an exhoneration, and Bush knew it at the time he made the announcement! Unbelievable! The way the science was distorted and spun to serve a particular political agenda is remeniscent of things like the Nazi idea of racial supremacy, and even Islamic (and other forms of) fundamentalism, the Christian crusades, etc. We should be smarter than that by now.
And Bush's nuclear policy is so bad that he should have been impeached on the same day he broke the nuclear Test Ban Treaty. Our outrage at that should have been so tremendous that we demand Bush be impeached, or face a nation-wide revolt to overthrow the government! His plan for bunker-buster nukes is completely implausible. I should have a link handy but I don't -- suffice it to say that these devices are impossible on basic physical principle. And his plan for chemical and biological agent-destroying nukes is also quixotic. A decent-sized nuke would destroy biological agents within about 10 feet of the bomb, and scatter the rest into the atmosphere -- a better *delivery mechanism* than anything else. yet Bush still values these initiatives more than he values a world more safe from nuclear holocaust.
So I voted for Kerry, for what it's worth. I am writing this after the nearly-inevitable reelection of Bush, and am honestly just beside myself with disbelief. I don't know what to do. Should I flee the country? No, that's a cowardly way out and won't solve anything. My oppinion of the average US citizen has just fallen to an all-time low. I guess the only thing to do is hope for and fight for education reform, so that maybe in a generation or two the U.S. won't be so overpopulated with hopelessly moronic brainwashees, if we're still around to see the day.
I've tried to purge the experiences from memory, so I remember the feeling and pattern, but not so much the particulars. One example I do remember (since I fought for a decent work-around so long) is automatic figure numbering. I don't remember the details here, and no longer have easy access to Office (Gentoo in front of me and Gentoo behind me!).... Aha! Automatic numbering doesn't work (or didn't for Office 97) inside frames. They do work from text boxes, but those don't have valuable features like text anchor, IIRC. Importing a figure into a table was hard enough, but using text boxes was even more difficult, and in the end was just unsatisfactory. This was a show-stopper because the documents I was working on were very heavy on figures.
I also found the HTML export to be broken. It produced VERY ugly and bloated HTML which still didn't format correctly. It looked bad in IE, but completely worthless in Mozilla. The tag ordering was incorrect, etc. This is certainly true of Office 97, and I think Office XP is even worse. It looks deceptively good in IE, but not in Mozilla so it's not actually producing HTML!
Also, Word doesn't handle EPS figures well at all. It can't generate previews and can't print on non-postscript printers. It does okay with WMF files, most of the time. But I've had numerous problems with these as well, mostly because they are hard to generate. Is the format another Microsoft proprietary thing? I also had tribulations with CGM. I haven't tried SVG. These days when using PowerPoint I convert to PNG at high resolution and suffer the performance penalty.
To be fair, I don't think OpenOffice deals with figures well either, and its mathematical typesetting is akward to use. MathType works well with Word, and it's relatively cheap compared with $500 Office. But not everyone has MathType installed, so it's not portable. And it doesn't work with PowerPoint completely (not inline anyway), which is otherwise Microsoft's best product. Now I'm getting way off topic.
What honestly concerns me the most with MS Office is the closed file formats. I don't think Dvorak makes a very eloquent case here, but at least he touches the issue. How many old documents do YOU have that are stored on obsolete media, in an obsolete proprietary format? I have a lot, and they're all useless to me. Word has decent support for reading older.doc formats, but are you absolutely positive that Microsoft will continue to grow and thrive and support old.doc formats? I'm not.
Also MS has long abandoned maintaining "foreign" import filters. Word can't read an OpenOffice document, but OpenOffice can read a Word document, not perfectly but much better than one should reasonably expect without MS code. (So OpenOffice is better than Office, case closed.) Anyway, it is not reasonable these days to be satisfied with having to store valuable data in file formats that are closed and proprietary. We didn't know better 20 years ago, but we know better now. If nothing else about Word strikes you as absolutely unsatisfactory, then this still should.
I didn't agree with everything he said either, but I can relate to the innumerable bugs in every version of Word that I've used (since it went GUI). I've always been surprised that people use it so much. I used it a lot during college. But over the past 10 years since, I've tried using the latest Word for various things probably 3 or 4 times, and each time I've encountered different show-stopping bugs that made me turn elsewhere (usually latex). This happens after about an hour of using it, and I would typically search the net for help and learn that it's a "well-known" bug, sorry, causing even textbook examples to fail. So I get the impression that Microsoft does less than 1 man-hour of quality-control on every version of Word they release. That's a little unfair to say, but only a little.
So I feel I can relate to Dvorak here. I'm sure that one can deal with Word if they make a career out of it after thoroughly digesting some book like O'Reilly's "Word 97 Annoyances", and learning all the work-arounds. But for the (effectively) novice user like me who will use another program after initial frustrations get too high, Word is just way too buggy to use.
I think that this would help Microsoft in the long term. But here I'm envisioning a future where an open OS, say Linux, has a dominant position on the desktop. If today's Win32 developers want any kind of longevity for their projects, they ought to be paying attention to which API are well-understood and simulated under wine and which ones aren't. IIRC Wine simulates only about 40% of the Windows API. That leaves a lot of potential sand traps for today's developers to avoid. So MS contributing to Wine isn't entirely insane.
However, I don't know how easily MS employees could sign-off on any of their contributions to the project.
Slashdot Mirror
More reasons:
Standards: It's amazing how many people send around Word, PowerPoint, and Excel documents. Why do they expect me to be able to read these? Do they expect me to cough up ~ $500 like they did (or didn't) for software I don't need other than to accomodate their whimsy? If they send me an OpenOffice document instead, there's no financial burden on me and it's an open standard which I am in some sense more entitled to use/interpret/read.
As for Firefox -- since this is now popular enough at my workplace, I basically don't need to test for compatability with IE (which is difficult to run in Linux), and I don't support it in my web projects. If it works in Firefox, Opera, and Konqueror (which almost tests Safari compatability), then it's good enough for me. I am also looking forward to more complete support for things like MathML, which will gradually make life much easier for me. As for IE support, I wish there was a web page that launches an ActiveX script that installs FireFox with little notification.
For an open-source project, the number of developers tends to increase with the number of users. I don't think it's a linear relationship, but it's certainly monotonic. And when talking about open-source, the distinction between developers and users gets wonderfully blurry.
Finally, I love the wide variety of open-source projects going on! I often find projects that are useful to me, and it seems like each year computing just becomes easier thanks to OSS. Better programming languages, more libraries, more complete hardware support, improved documentation, etc. The more people become aware of open source, the more they will get involved in it. This is of direct benefit to me, and everyone else too!
I think your underlying assumption is that if all drivers act in their own best interest, the total efficiency of the system maximizes (I mean that the average driver's trip time minimizes).
It's not obvious to me that this assumption is true. It may well be the case that the best overall traffic pattern involves sub-optimal routes for some drivers.
I think FreeCiv is a leading example of how OSS can enable collaborative game balancing and development. For a moment, don't think about the software at all. Just look at the game mechanics, and how FreeCiv has been fine-tuned by a community of interested players. That's really kind of new! The typical game development model these days is that the game company works on balancing totally in-house, and might to a limited extent actually listen to the players. The result is often frustrating to fans of the game who would like to try out their own ideas.
FreeCiv definately leads the pack with a large active community and a long history of development. I don't think you can count nethack in this class because, AFAIK, its core development team has remained an exclusive group.
Yeah, that quote from Roper hit a nerve with me too.
It's too bad that the term "RPG" now has this dual meaning. The two definitions are worlds apart. It means either "computerized stats-based avatar-building game", or "role-playing game". The former refers to strategy games like Diablo where players try to perfect chcaracter builds and collect ever-more fantastically rare and powerful items to either compete against each other or to make the game easier and easier.
That, obviously, has nothing whatsoever to do with role-playing where the objectives are story telling, creating an atmosphere, and developing a believable personality. I'll argue that Neverwinter Nights was the first computer game which aspired to be an RPG, as it was the first game to really allow for a DM.
Thanks for the tips! I'm looking into firestarter and transcode wrappers right now.
You've made a lot of great points here IMO.
I wanted to add a couple ideas:
On the pro-Linux side, I had a night-and-day experience upgrading one of my computers to a new MB/CPU/video card. It dual-booted to Windows 98 and Linux. I wanted to keep Win98 because it's a computer that my young daughter uses to play a bunch of kids' games. Upon first booting up, I went into Linux. Perfection! Zero work required. I was upgrading from an really old (RIVA TNT2) NVidia card to an old NVidia (GeForce3 Ti) card, which made life easier for Linux.
Then I tried booting into Windows. And tried, and tried. For a long time I had to turn off "plug-and-play OS" in the BIOS because Windows kept all the incorrect BIOS drivers. I even had to fight to get the CD-ROM working so that I could give it the Windows install CD to loads dozens of drivers off of. Every single piece of hardware required 1-4 reboots to get working, independantly. Honestly I struggled with this for about a week, a couple hours a day. Among other things I had to re-assign the CD-ROM drive letter manually three times during the week, each time so that it could find the $@@#%&* Windows CD. Unbelievable! Admitedly Windows 98 is old and barely supported these days, so to be fair I should compare it to, say, RedHat 5.2 or something like that. But even Windows XP is not a piece of cake when dealing with massive changes to hardware.
Okay, my other comment is that the newbie's experience with Linux (and I mean Linux newbie, not computer newbie) varies a whole lot depending on what hardware they are using, and what they are wanting it to do.
For example, I helped a friend install Gentoo Linux on an Athlon 64 laptop, going for the full 64-bit OS and applications. It works now, but it was not easy. It took a while to get the broadcom wireless card working with ndiswrapper, get accellerated X with his ATI video card, etc.
You can't really blame Linux developers for difficulties supporting the latest hardware. Microsoft doesn't have nearly as hard a time since they are so popular that all hardware manufacturers write their own Windows drivers. Not only do Linux users have to write their own drivers, they often can't get the specs and have to reverse engineer it. It's a major testament to the sheer power of the open-source model that so much hardware is supported in Linux, often with better reliability than the Windows drivers!
Then there's a lot of powerful Linux software with a steep learning curve. MythTV is kinda hard to set up right. Doing firewalls well with ipchains requires a lot of (worthwhile) reading. Working with video using transcode is also time-consuming and involves lots of trial and error.
So for the typical office computer user using well-established hardware and a nice easy distro like Mandrake or Knoppix, Linux can be substantially easier to use and learn than Windows. But the power-user or hobbyist will find no end to the cool things that Linux can do with a lot of hard work. And someone running Linux should not expect to buy a GPS device + MS Streets software and literally plug and play!
While we're bashing Mac mice, I still vividly recall the negitive impression the circular iMac mouse had on me back in '97 or so. One of the campus public computer labs had some nice scanners that I would use every few months, and they had just updated their Macs to the colorful, translucent iMacs. There was no indentation for my fingers on the one button (I too feel like I'm wearing a mitten on a Mac) and the mouse was perfecly axialy symmetric, and so there was no way to feel the orientation of the mouse in my hand. Every single time I went for the mouse to click on something, the cursor would go off in some random direction and it would take me a few seconds to re-orient myself!
I always hear that Apple holds aloft the banner of greater usability and productivity, but I just can't take seriously any company that would let a mistake like that out the door. Had no one even tested the mouse for three minutes before mass-producing it?
Okay, that was a long time ago. But it was so disconcerting that I still think of it whenever I hear about Mac mice.
From what I hear, the argument for the one-button mouse is that some small fraction of new users will experience confusion and frustration with something more complicated. But surely that confusion would not last too long? So Apple really wants the majority of their users to buy another mouse to replace their functionally-challenged one? What about public computer facilities where the admins generally go with the original equipment? I'm glad they may start making two-button mice, but that scroll wheel is very handy, too. I don't personally like the side buttons because my thumb and pinky are used to grip the mouse and I find that I click those buttons by accident a lot. Also, the extra functionality of the side buttons is very limited in scope, and can be conveniently replaced with things like mouse gestures.
Oh ... Heh ... Sheepish now. My wife does our taxes anyway -- and a good thing too!
I agree, www.silentpcreview.com does have a lot of good advice. But like any review site, they only cover a small fraction of the current hardware available that is marketed to the quiet PC crowd.
I spent some time and money last month going through that site and working to quiet my PC down. I set up a new PC to be always-on in a corner of my dining room. The location was dictated by cable access and wi-fi reception. (It's running Gentoo Linux and is serving as a gateway/firewall, records radio shows, web server, etc.) The dining room has hard floors, and the PC noise was absolutely unbearable at first. Now it's just about as loud as the 700W APC UPS it's on, which is noticeable in that room but not at all obnoxious.
* case: Evercase GC4292: inexpensive steel case with no P/S and can handle a 120mm case fan.
* P/S: 300W Seasonic supersilencer -- has a very quiet fan, I really like it.
* CPU fan: Arctic Cooling Copper Silent 2 TC -- I'm very happy with this too, barely audible with case open.
* VGA fan: Zalman VF700-ALCU Ultra quiet fan -- Another winner, though the RAM heatsinks were unnecessary (I'm not overclocking) and didn't even all fit without compromising the main heatsink's contact with the chipset.
* I mounted the HDD with rubber grommets. The case's HDD bay supports these but I had a hard time finding the kit to go with it. I'm still looking for more grommets. I believe that this helped substantially, but I don't have test equipment and I swapped hard drives at the same time. The old drive without grommets was a prominent noise source, and now I don't hear it except when the head is seeking. I don't remember the brand/model, but I think it's a Maxtor 5200RPM, 200G drive.
* I am still struggling with a quiet case fan. This is now the only component I can really hear with the case closed up. I tried two Antec variable-speed fans with thermistors on them, one 90mm and the other 120mm. I'm currently using the 90mm fan because the 120mm fan has a louder and more annoying pitched hum. I like the fan speed control idea in this article.
I also bought the Antec noisekiller kit in a whim, but it was difficult to use and I don't think it really helped any. I was hoping for drive mounting screws and grommets, but they were not included.
While there is a lot about the US tax laws that I find ridiculous, I at least have some appreciation for the difficult problem of fair taxation. We have hundreds of millions of people trying to save on their taxes, and if any loopholes exist, they will be found and exploited. While I would love a simpler system, there is balance. If it's too simple then it's easy to abuse, and if it's too complex the same could be true. An abusable system is worse because you'll either pay more than your share, or you'll pay a tax preparer/crook to find sneaky ways to save money. You want a system that doesn't demand a paid tax preparer, and thus you want a more airtight one.
We are already in this bad situation somewhat, but I think that while it could easily be worse, it's not easy to make it much better than it is now.
I also don't think that the tax preperation lobbyists/donors have more influence than the tax-payers who want an easier system. We outnumber them. So the argument that the tax-preparers force the IRS to keep them in business doesn't make sense.
One of my pet peeves is the tax bracketing system. One can save a fistful by managing to squeeze into a lower bracket. This encourages abuse. To make taxation a continuous, monatonic function of earnings requires no more than grade-school mathematics. But it would require just slightly more math than the present system: multiplication followed by subtraction instead of straight multiplication. So the IRS deems that the subtraction pushes the requirements into the lair of advanced mathematics, and thus the idea doesn't make the simplicity cut.
It's our own fault, really. We ask for simpler tax law and this is what we get.
I can second both of these games. Scotland Yard is such an original game concept! Lots of fun.
Empire builder and follow-up crayon rail games are also great. I think Eurorails is my favorite, since the games often come out unexpectedly close with the winner only a turn or two ahead of the rest even though their tracks and strategy are completely different from other players. I've always wanted to computerize these rail games but haven't found the time yet. The main reason would be to play with friends and family that are living 3000 miles away. And partly this would be fun just to collect statistics to refine track-building strategies and work out good distributions of demand cards. Also it would be a fun game to design an AI for.
"The most powerful sort of aptitude is a consuming interest in some question, and such interests are often acquired tastes."
I heard one study recently demonstrating that in grade school, girls do better in science and show more interest then boys. In high school, men edge out the women, in college (for the hard sciences anyway) it's more like 3 to 1 men-to-women, and in grad scool it's more like 10-to-1. One interpretation of this is that innate interest or capability is not the problem, but social pressures are. Is this a problem? Yes! Even if for no other reason than that science needs diversity of thought to make the most progress.
A non-PC, unpublished, didn't-officially-happen study I heard about was histogramming grade distributions for the colegate physics 100-level course series (a couple thousand students). The men's grades fell into a neat bell curve with maybe a B average. The women's grades were a remarkably flat, uniform distribution from 0 to 100 percent (where F was about 50 or 60) -- remarkably distinct shape from the men. I'm not sure what conclusion to draw from that, but in my oppinion it's a serious problem.
No, not everyone should excell at sciences, and it's certainly true that the hypothetical average woman has different strengths and weaknesses than the average man. But the 100-level physics courses should not require extraordinary ability to score well in (just hard work at worst), and I have no reason to expect that women are naturally worse than men at it either. I also think that someone's aptitude for sciences depends primarily on their own self-expectations.
My first reaction to this was that this MUST be a hoax. It's just the kind of thing my friends and I would have joked about in high school. For the record, we also came up with the McCow which is a rectangular (well cuboid) bovine that's stackable. These occasionally get loaded onto trolley cars and driven through an astroturf pasture for the visitors. But that's another story...
My second reaction is that for this to be in any way scientific, we need free access to the source code. It sounds like we're presented with basically one alrogithm that converts a song into 2 or 3 numbers, and these are compared with the distribution of other known-to-be-popular songs. Well, what happens when these algorithms change? How stable are the results against using different algorithms or different constants or different datasets of known-good music? The lack of this kind of in-depth analysis (which would have to be done by an independant party to have any validity) calls into question the whole technique.
I still don't like the idea of a computer algorithm helping to determine what future songs get airplay, but maybe it's not any worse then our present situation. Today we have "top-40" stations with play lists less than 20 songs deep, and big-wigs that listen for the hook in the first 15 (IIRC) seconds of a song.
Is there any open-source backup software for Windows? I have a little Perl script that I've been using for my laptop running Linux that uses tar for compressed, incremental, selective backups onto a public backup SMB share (or any mounted directory) after gpg encryption. I can "port" it to cygwin very easily and may use that if I don't find anything friendlier.
Oh, nice! I like the gallery program at first glance. I'll investigate it more soon.
I wrote my own pictures -> HTML gallery program a few years ago, using Perl and ImageMagick (via the Image::Magick Perl module). But it was never quite advanced enough to release it -- just a couple hundred lines of code (which accomplishes a lot in Perl). I still use it from time to time, often just to make thumbnail images for other purposes.
I usually bring my laptop, and show pictures on that. The screen is not large, but at 1600x1200 resolution the pictures look VERY nice! Admitedly not more than 3 or 4 people can really gather around a laptop, but it does the pictures justice. A can't imagine that a TV would be acceptable unless I only had a 1/4 MP camera.
I have to mention software, too. gqview. I haven't found any Windows software that even comes close to gqview. I suppose the best I've found so far for Windows is Irfanview, but compared to gqview it's still awful. Now I read somewhere that gqview works under cygwin. For my Windows box, I'll try that next.
Also, as for taking good pictures, I'm loving my Canon digital Rebel. I heard that professional photographers take around 100-1000 pictures for every one they like. I'm not a pro, but my standards are much lower. So I think my ratio is about the same.
Digital allows me to take pictures indescriminantly, and try all kinds of artistic shots that I wouldn't dare waste film on. I learn A LOT from these, and sometimes I get a gem. On a typical day hike I'll take a couple hundred pictures. Of those, I'll delete about a third (near-duplicates or badly flawed), really like a dozen, and might find 1 or 2 that are (for me) very good photos. I have a couple scripts that help me sort pictures into directories (using hard links), and these are ones I would show to people if I have the time.
I didn't like the article at all. I think there are several factual errors, trumping all the problems with the flash presentation.
Is a larger sensor really necessarily better than a smaller one? I don't think so, but it definately places different requirements on the optics. Perhaps megapixel-for-megapixel the larger sensors are sharper because of bleeding from one pixel to another? Anyway, it's the combination of optics and sensor that determines image quality (and I suppose post-processing, etc). DSLRs can have good-quality optics, but the larger sensor requires larger lenses, and that makes them more expensive for the same amount of quality.
Depth of field -- he claims that no camera can have the same depth of field as the human eye. Again, just not true. We are constantly focusing on things near and far, but at a given focus, we have shallow depth of field. An SLR camera gives you the most control over depth of field, and is generally much more narrow than a compact digital. With an artistic eye, this can be a great advantage of the DSLR.
I'll read the rest later. He does make some good points. But I was not impressed initially.
I hope that he mentions the shutter lag times and focusing time. Those aspects are very important for quick candid shots, and the DSLR's really beat out the compacts here.
Agreed 100%!
All Blizzard's other games cost about $50 new, but there was no subscription fee and you could download a free demo. That sales model is fine with me. They talk about supporting the online community and adding more content, but they've done that with all their other games too, again w/o the subscription fees.
If they only charged the monthly fee, that's like a close-to-free ($15) limited-time demo, and if you like the game and play it for more than 4 months (Most Blizzard games maintain their appeal for years) then you've already spent more than $50 on it.
Also, I like online games that I can play with my wife. Again, most games don't really require me to buy two copies for LAN play. But now I'm looking at $100 PLUS $30 per month. That's nearly as much as a broadband connection (which also is usually free installation)!
I'm sorry. This game REALLY sounds awesome, and I would LOVE to play it. I'm totally into these kinds of games. But the cost is just way too high. That kind of monthly fee is much too large a fraction of my monthly savings.
Also keep in mind that Blizzard sold 250,000 copies on the first day. Suppose they've sold 1 million by now. That's 50M already, plus 15M per month. I think that's far more than enough to support the 80+ servers and staff. Blizzard is getting greedy.
The article defends this point in several ways:
First, after several tens of years, the composition of fuel rods changes significantly -- the shorter-lived components will decay and the waste will generate far less heat. The ideal storage environment changes substantially then.
Also our current waste-management techology is immature, and not proven to be good enough. But a few new developments are on the horizon.
Future technology is likely to make fuel reprocessing more economic (and I think he did this without even mentioning breeder reactors).
Finally, the Yucca mountain storage facility is gridlocked in politics and thus not a realistic short-term option.
Then the author addresses your question, and suggests using our present temporary solution of casques but upgrading it to a centralized facility that can be hardened against terrorist abuses. It's not clear to me that this is the best way to go but it's obvious that we need an *immediate* improvement over what we're doing now, and that we really want to consider a temporary storage technology that's good for 100-200 years, not necsessarily 100-200 thousand years.
I was thinking about this very recently. What most distinguishes journalism from science?
In a good scientific article, the author is often her own best critic. The paper will first establish the facts: precisely how the experiment was done, and what the results were. That's usually the easiest part. The great majority of time spent is in figuring out just how conclusively the data confirms the thesis. This involves looking for alternate explanations for the data, and guessing at any possible systematic bias.
After reading a good scientific article, you should come away with an understanding of what confidence there is in the conclusions. This study of uncertainty and systematics is really the most important part -- it's perhaps the defining characteristic of science! And this element is precisely what's lacking in journalism.
Every scientist will have a personal bias going into a study, but does their best to quench that bias with a strong and healthy skepticism. And a reader or journalist can, by reading multiple articles addressing a particular issue (independant confirmation is another cornerstone of science) gain an more precise and less-biased oppinion.
So, to be unbiased, a journalist doesn't have to be "ballanced" but they should compile as many differing oppinions as possible, and judge them by the strength of their evidence. Any report of science would be incomplete/broken/useless without these three elements: discussing the body of evidence, reporting the (statistical) uncertainities, and investigating for possible bias.
The articles refered to in the story do mention the strong personal biases of the fringe scientists. But it would be even more convincing to criticise their research, not their politics.
Journalists often have the hard job of distilling science for the masses. But I think they go way too far, presenting a collection of scientist's oppinions rather than presenting scientific evidence. If the focus were on the data and not the oppinions, the personal biases of the researchers and journalists would be largely removed.
To conclude, journalists absolutely must preserve:
1) The evidence -- most importantly what experiments were done.
2) Uncertainty -- how convincing is the evidence?
3) Bias -- how susceptable are the experiments to bias?
4) Independant confirmations -- not how many other scientists agree at a convention, but how many fundamentally-different experiments support the same hypothesis?
I was actually considering voting for Badnarik, on the premise that Instant Run-Off voting is actually a vitally important election reform for the US. The Republican party so far has the least incentive for establishing it. More votes for Badnarik would help with this. And I live in New York, so "my" electoral votes will go to Kerry either way.
But then I figured that the popular vote is also somewhat meaningful, and I should really only throw away my vote if I were relatively neutral between Kerry and Bush. I don't like either one really, but I think that Kerry would make a far less-awful president than Bush has been.
As for Kerry, although his anti-Bush statements have been on target and valid, his own oppinions on what to do in Iraq are unrealistic at best, pandering at worst. He might sound good, but I feel that many of his arguments have no substance. I'm also disturbed by Kerry's incessant trumpeting of Cheney's daughter. I don't know of any "real people" who would make such a big deal out of such a trivial fact, going out of their way at every turn to make polite-sounding comments that are thinly-disguised attacks like these. It just disgusts me. No decent human being would do that. Kerry did. Sigh.
But all of these issues that might in a normal elecetion swing my vote are suddenly made trivial by the outrageous actions of Bush. His war on Iraq was based largely on evidence of nuclear weapons, which we later find out was such weak evidence that it was more nearly an exhoneration, and Bush knew it at the time he made the announcement! Unbelievable! The way the science was distorted and spun to serve a particular political agenda is remeniscent of things like the Nazi idea of racial supremacy, and even Islamic (and other forms of) fundamentalism, the Christian crusades, etc. We should be smarter than that by now.
And Bush's nuclear policy is so bad that he should have been impeached on the same day he broke the nuclear Test Ban Treaty. Our outrage at that should have been so tremendous that we demand Bush be impeached, or face a nation-wide revolt to overthrow the government! His plan for bunker-buster nukes is completely implausible. I should have a link handy but I don't -- suffice it to say that these devices are impossible on basic physical principle. And his plan for chemical and biological agent-destroying nukes is also quixotic. A decent-sized nuke would destroy biological agents within about 10 feet of the bomb, and scatter the rest into the atmosphere -- a better *delivery mechanism* than anything else. yet Bush still values these initiatives more than he values a world more safe from nuclear holocaust.
So I voted for Kerry, for what it's worth. I am writing this after the nearly-inevitable reelection of Bush, and am honestly just beside myself with disbelief. I don't know what to do. Should I flee the country? No, that's a cowardly way out and won't solve anything. My oppinion of the average US citizen has just fallen to an all-time low. I guess the only thing to do is hope for and fight for education reform, so that maybe in a generation or two the U.S. won't be so overpopulated with hopelessly moronic brainwashees, if we're still around to see the day.
I've tried to purge the experiences from memory, so I remember the feeling and pattern, but not so much the particulars. One example I do remember (since I fought for a decent work-around so long) is automatic figure numbering. I don't remember the details here, and no longer have easy access to Office (Gentoo in front of me and Gentoo behind me!). ... Aha! Automatic numbering doesn't work (or didn't for Office 97) inside frames. They do work from text boxes, but those don't have valuable features like text anchor, IIRC. Importing a figure into a table was hard enough, but using text boxes was even more difficult, and in the end was just unsatisfactory. This was a show-stopper because the documents I was working on were very heavy on figures.
.doc formats, but are you absolutely positive that Microsoft will continue to grow and thrive and support old .doc formats? I'm not.
I also found the HTML export to be broken. It produced VERY ugly and bloated HTML which still didn't format correctly. It looked bad in IE, but completely worthless in Mozilla. The tag ordering was incorrect, etc. This is certainly true of Office 97, and I think Office XP is even worse. It looks deceptively good in IE, but not in Mozilla so it's not actually producing HTML!
Also, Word doesn't handle EPS figures well at all. It can't generate previews and can't print on non-postscript printers. It does okay with WMF files, most of the time. But I've had numerous problems with these as well, mostly because they are hard to generate. Is the format another Microsoft proprietary thing? I also had tribulations with CGM. I haven't tried SVG. These days when using PowerPoint I convert to PNG at high resolution and suffer the performance penalty.
To be fair, I don't think OpenOffice deals with figures well either, and its mathematical typesetting is akward to use. MathType works well with Word, and it's relatively cheap compared with $500 Office. But not everyone has MathType installed, so it's not portable. And it doesn't work with PowerPoint completely (not inline anyway), which is otherwise Microsoft's best product. Now I'm getting way off topic.
What honestly concerns me the most with MS Office is the closed file formats. I don't think Dvorak makes a very eloquent case here, but at least he touches the issue. How many old documents do YOU have that are stored on obsolete media, in an obsolete proprietary format? I have a lot, and they're all useless to me. Word has decent support for reading older
Also MS has long abandoned maintaining "foreign" import filters. Word can't read an OpenOffice document, but OpenOffice can read a Word document, not perfectly but much better than one should reasonably expect without MS code. (So OpenOffice is better than Office, case closed.) Anyway, it is not reasonable these days to be satisfied with having to store valuable data in file formats that are closed and proprietary. We didn't know better 20 years ago, but we know better now. If nothing else about Word strikes you as absolutely unsatisfactory, then this still should.
I didn't agree with everything he said either, but I can relate to the innumerable bugs in every version of Word that I've used (since it went GUI). I've always been surprised that people use it so much. I used it a lot during college. But over the past 10 years since, I've tried using the latest Word for various things probably 3 or 4 times, and each time I've encountered different show-stopping bugs that made me turn elsewhere (usually latex). This happens after about an hour of using it, and I would typically search the net for help and learn that it's a "well-known" bug, sorry, causing even textbook examples to fail. So I get the impression that Microsoft does less than 1 man-hour of quality-control on every version of Word they release. That's a little unfair to say, but only a little.
So I feel I can relate to Dvorak here. I'm sure that one can deal with Word if they make a career out of it after thoroughly digesting some book like O'Reilly's "Word 97 Annoyances", and learning all the work-arounds. But for the (effectively) novice user like me who will use another program after initial frustrations get too high, Word is just way too buggy to use.
I think that this would help Microsoft in the long term. But here I'm envisioning a future where an open OS, say Linux, has a dominant position on the desktop. If today's Win32 developers want any kind of longevity for their projects, they ought to be paying attention to which API are well-understood and simulated under wine and which ones aren't. IIRC Wine simulates only about 40% of the Windows API. That leaves a lot of potential sand traps for today's developers to avoid. So MS contributing to Wine isn't entirely insane.
However, I don't know how easily MS employees could sign-off on any of their contributions to the project.