The fact that the language under critque was FORTRAN, unfortunately today, obscures the underlying truths they were discussing.
I disagree. That book rocks; every software geek should read it. Yes, some of the examples are non-interesting because they are too FORTRAN-related. But lots of the other examples are timeless classics. The discussion of floating-point math is a gem. "Floating point math is like moving piles of sand around; every time you do it you lose a little sand and pick up a little dirt." (Quote from memory so it might not be exact)
And by the way, about half the examples are in PL/1, not FORTRAN.
One of my favorite features in Evolution is the safe preview: when you look at an email message, it renders the message in all ways that do not involve hitting a server. So an HTML message with bold, italics, colors, pictures, etc. will display correctly... except for anything that would touch a server.
Why is this important? Because spammers make special URLs that encode your email address, and their servers use the special URLs to track you. If you even preview the message in Outlook, or in Mozilla Mail, their server gets a hit and they know your email address is a good one.
There is a command on a menu, "Load Images", that will go ahead and put hits on servers and render the message completely. You can use this for email from sources you trust. (It ought to be a toolbar button, but it isn't yet in Evolution.)
If Thunderbird doesn't already have this, I ought to file a bug.
You said that "All of the Java platform is available on a free license for open-source developments", but it looks to me like Sun is locking down anything new under the SCSL. SCSL isn't a free license; the Kaffe guys say so, and I agree with them.
So, are you saying that the SCSL is a free license, or are you saying that Sun is releasing new Java stuff under some other license, or what?
Can you give me a link to some web page or email or something that says that I'm wrong and Sun released the latest Java specs under a free license?
steveha
Re:OS/2 is a separate issue?
on
Nat Demos Dashboard
·
· Score: 2, Interesting
You are mistaken about Microsoft and OS/2. Microsoft really believed that OS/2 was the future. I worked at Microsoft in the early 1990's and everything was all OS/2, OS/2, OS/2; developers all had OS/2 computers for development work, the computers in the library ran OS/2, all Microsoft applications had OS/2 versions available, etc.
Customers voted with their dollars, and they voted for Windows rather than OS/2. I believe this was due mainly to the fact that Windows had a much easier migration path: if you had several DOS apps that you needed, you could run them all in Windows, versus running one at a time in the compatibility box under OS/2 and possibly crashing your computer. (Yes, later OS/2 versions were better, but that was after Windows had already won and Microsoft was already gone.) Other issues were that Windows ran much better on the computers that people had back then, and that Windows cost less than OS/2.
So, once Microsoft figured out that the customers wanted Windows and didn't want OS/2, Microsoft made the famous deal with IBM where IBM got OS/2 and Microsoft kept Windows. Microsoft didn't betray any OS/2 users, because IBM was there to support those OS/2 users.
In summary, Microsoft didn't have some cynical bait-and-switch plan, because internally Microsoft was pushing OS/2 right up until the famous "divorce" from IBM. And Microsoft didn't "pull the rug out" because IBM was fully supporting OS/2. It's not Microsoft's fault if IBM wasn't able to take over the world with OS/2.
Microsoft does have some things to answer for, but this really isn't one of them.
FACT: All of the Java platform is available on a free license for open-source developments, including the test suites. This is what the Kaffe people use. Nothing comparable exists for Dotnet whatsoever.
I wish things were as rosy as you make them sound. Kaffe is stuck in Java 1.x compatibility, because Sun Microsystems is keeping the Java 2.x specifications unfree. See section 2.3.1.5 , "Why is (some) free software not implementing Java2?", of the Debian Java FAQ.
I have a classic, wonderful monitor: an SGI 1600sw, with 1600x1024 resolution. I only run it in its native resolution. My fonts are large and beautiful.
0) Use TrueType fonts.
Make sure your X11 setup is all correct for scaled fonts, especially TrueType fonts, and then get some good ones. Go everywhere and make sure you are using your good TrueType fonts. My GNOME preference fonts are all TrueType, plus my web browsers. The GNOME 2.x dialog for this is Applications / Desktop Preferences / Font. If you have an LCD flat panel display, be sure to check the box that says "Subpixel smoothing (LCDs)". For a CRT monitor, I suggest you check the "Best shapes" box.
1) Grow your fonts.
Go into XF86Config (or, in Debian, XF86Config-4 if you are using a recent version of XFree86). Find the part where it describes the monitor. There should be a DisplaySize line describing how big your monitor is, in millimeters. If the line is not there, search the web for specs on your monitor, or just measure it, and add the DisplaySize line. For the 1600sw:
DisplaySize 369.4 236.4
Now we want to lie to X11 about the size of our monitor, and say it's smaller than it really is. I want fonts 150% the usual size, so I multiply each number by 100/150 (i.e. 2/3 or 0.66666).
DisplaySize 246.3 157.6 # lie to get 150% font zoom (165 dpi)
# DisplaySize 369.4 236.4 # correct: 1600x1024 at 110dpi
Note that I like to leave comments about what the heck I'm doing here and why.
Now, X11 thinks my monitor is 165 DPI, instead of the real 110 DPI. When an application asks X11 to display 12 point text, X11 scales the TrueType font accordingly. I get automatic, across-the-board font zoom.
Peeve: there ought to be an X11 setting for this. You ought to be able to specify a zoom level, say 150%, and have X11 honor it without bastardizing the monitor size. If I can't have a zoom level setting, at least let us specify the DPI as a DPI number, instead of as the number of millimeters our monitors are!
2) Grow your web fonts too
Now your other big problem will be web sites that hard-code sizes. Even with 150% zoom, you really don't want 6-point fonts. The "minimum font size" setting in Mozilla hasn't worked well for me when I tried it in the past. You can specify a horrid large font size in the prefs, but then when you print a page, it prints huge too!
The solution is to use a cascading style sheet (.css) file. Go to your ~/.mozilla/default/<something>/chrome directory, and edit a file called userContent.css. (Be sure to check out the example files that Mozilla leaves there for you, while you are there!)
These lines mean: only for display on the screen (not while printing!), set the font size to 28 pixels height, and the line height to 30 pixels height. The "!important" part means you insist, even if the web page specifies a smaller size.
Now revel in the easier-to-read text.
You still have problems. Web designers who lay out pages with tiny fonts didn't expect their fonts to be forced huge, so the page won't look right; it might look downright ugly. And this fix does nothing to help when the webmaster specified a column width in pixels, so you may find a column that was intended to be over half your screen width is actually only three inches wide! (Thus you have big, easy-to-read text in a skinny very tall column, and you have to scroll the page a lot to read it.)
You also may find some text-entry forms that are 6 points tall, but the text you are typing into them is still huge, so you can't really read it. I ought to figure out what preference sets minimum text-entry box size.
Anyone with more useful tips, please share them!
P.S. Slashdot would not let me include the lines from my config files
I have a Tungsten T and I love it. When the newer Palm devices came out, I said to myself, "I wish Palm would come out with a device just like the Tungsten T but with the new tranflective screen". Here's the T2, and it fits the bill.
Double the memory is nice, the transflective screen is great, and the rest is almost purely unchanged. That spells a winner in my book.
I'm nervous about Graffiti 2; I like Graffiti as it is. But Palm lost a lawsuit over Graffiti and they don't dare ship it in new models, so I don't blame them for the change, and perhaps Graffiti 2 will be a good thing anyway.
It also sounds like they have sweetened the software bundle, also a good thing.
Some posters have commented that this is "nothing new". I say it's exactly what I wanted. A Tungsten T is a sweet little PDA, and the T2 should be better still.
Your wider point, that there are choices in the Palm market, is spot-on. If Palm were to die and take their proxy with them, you would be able to get some other browser for your palm.
...I don't see any positive benefit to purchasing the license unless they intend to go after each business individually in court.
However, they are hinting that they will do exactly that. They sent letters to 1500 companies they believed were running Linux, saying that the companies themselves might be liable (i.e. that SCO might sue them later).
In practice, it would be a very big job to go after a bunch of companies, each in turn, one at a time. But that's the threat they are using to encourage companies to buy their licenses.
I'll second the idea of bicycling to work. I was working far too many hours at a desk, and I started riding a bike to work; it made a difference for me.
It took me 20 minutes to drive to that job, or about 50 minutes to ride my bike to that job. So biking really only cost me a net 30 minutes each way. I spent 60 minutes to get 100 minutes of exercise; such a deal.
I was riding 10 miles each way to work. If you live too far away from work, do a bus/bike or car/bike combination: ride partway on the bus or car, then go the rest of the way on your bike. (I'm assuming that the buses near you have bike racks on the front, like the ones near me. If not, maybe you can arrange to store a bike somewhere that the bus can take you to.)
The secret of happy bike commuting is in the equipment. You need good clothing, including rain gear (jacket, helmet cover, shoe covers, etc.). You also need a good bike, with fenders and a cargo rack.
Stash some clothes at work, so you don't have to haul lots of clothing every day when you ride.
Bicycling is a good aerobic exercise, great for burning calories. It will work even better if you can combine it with weight lifting or some other exercise that works your arms, back, etc. and not just your legs and cardiovascular system.
Red Hat could box up a single disk. This disk would install a Red Hat base system, which would then be able to use the Internet to download the rest using Up2Date.
Note that this one disk, if done correctly, could be distributed pretty much forever. Currently, when a new version of Red Hat comes out, a bunch of the old version goes into the trash. The one-disk installer wouldn't have that problem.
Note also that Debian works this way now too. The PGI installer is one CD (and that one only about 1/6th full) but it installs the latest and greatest versions of all your packages, from the Internet. There is even a bootable Linux recovery disk that can, as one of its features, install a Debian base system that you could bootstrap into a full Debian system.
Anyway, a single installer CD in a shrinkwrap box is all they need, as long as there is info on where to order a set of the latest packages on CDs for a CD install. You should also be able to buy those CD sets from CheapBytes or similar distributors.
But remember the "Freedom Zero" debate. Tim O'Reilly commented that the first freedom of free software, and he called it Freedom Zero, is the freedom to choose which license you use to release it. RMS rejected this concept, calling it "Powerplay Zero", saying in effect that only the GPL is acceptable.
ESR then challenged RMS publicly: if you had the power to require all software to be GPL software, would you do it? RMS did not answer. A "yes" answer would be consistent with his other public statements, however.
So I don't know if we can prove that one of RMS's goals is the eradication of commercial software, but his public statements are consistent with that position. A relevant quote from the "Powerplay Zero" essay:
Proprietary software is an exercise of power -- it harms the users by denying their freedom.
By the way, RMS has never been opposed to people charging money for software; he just insists that the software must be under the GPL (and thus come with source code, be subject to forks, etc.). This, in the real world, means that there is a (fairly small) upper bound on how much one can charge for the software.
[IBM] can hold those patents (and charge others for the use of them for non-GPL'd applications) and GPL code based on those patents.
Yes, exactly correct. This is why Linux has RCU and BSD doesn't. IBM is willing to license the patents for free for GPL code, but they still want to charge license fees for use in proprietary code.
Releasing something under BSD is effectively placing it in the public domain.
By the way, you sometimes see people claiming that "if there weren't any copyrights there would be no need for GPL" or some such. Not so. GPL uses copyright law to prevent people from taking free projects proprietary; BSD lets you do anything you want, including taking a free project, hacking it up slightly, and releasing it as a proprietary product.
It would be a hefty task for the filesystem to transparently convert/album into the appropriate ID3 (or whatever it is for ogg) format while allowing this functionality for every conceivable format (mp3, ogg, wmf, ra,...).
Plugins. Customizable plugins that run in user space. The article talked about this a little bit.
Just on general principle, I prefer my data inside the file and not left with the filesystem. The MP3 metadata example, to me, is like Windows file extensions on HGH.
I'm with you -- I like self-contained file formats.
But I don't think he was proposing that you not use Ogg tags or MP3 tags; he was talking about the filesystem abstracting the tags. If you changed "Stagnation.ogg/album" to the string "Trespass", then the filesystem abstraction layer should update the Ogg "album" tag inside the file to be "Trespass".
The key benefit here is that you would not need some wacky command-line utility program to let you view and change tags on Ogg files. You could just use the shell. In bash:
for ii in *.ogg; do echo "Trespass" > $ii/album; done
Note that this same one-liner would work if you were in a directory with MP3 files, and you changed "*.ogg" to "*.mp3". Currently, you need to run vorbiscomment for your Ogg files, and mp3info for MP3 files. (I just checked, and sure enough, they take different arguments to do the same operations.)
Personally, I'd like to see a standard metadata portable XML format for legacy systems. People talk about copying a file from a rich metadata filesystem and having new files like.attributes created on the target legacy system; I'd be happier if just one big XML file could be created with the same name as the original file.
Suppose you backup server//rich onto server//legacy, and then you want to restore some files from//legacy to//rich. If all the metadata was stored in a big XML file, then when you copy the file from//legacy to//rich you restore all the metadata; you wouldn't accidentally slice off attributes by forgetting to copy one or more rich attributes files.
You could do most of the fancy tricks of the rich metadata filesystem on a legacy filesystem that used the big XML file to store the rich metadata. And as long as the legacy system is just smart enough to look at the main data part of the XML and leave the metadata tags alone, you could still modify the file with sed, awk, perl or whatever, and then copy the big XML file onto your rich metadata filesystem and still not lose any rich metadata.
Note also that the big XML file could be used to deal with existing rich metadata systems, like resource forks from Macintosh filesystems, or multiple data streams from NTFS files.
To make me happy, please add these features to the toolbars:
Next to the delete button, a "spam" button. This would delete the mail message, but also flag the message as a spam. You should be able to customize this to run your local spam filter training script.
In the message display toolbar, there should be a button that loads images. By default, Evolution does not render any HTML that puts a hit on a server; this is a great feature, because spammers do use tricky URLs that encode your email address. When their server gets a hit from the URL with your address encoded, they know your email address is good, and you get more spam. Currently, the "Load Images" command is in a menu, and is not available with one click. "Load images" should also be in the context menu (the "right-click" menu). For extra credit, Evolution should remember that you already loaded the images, and if you go back again it should be able to get the images from the browser cache, not have to hit the server again as it currently does.
There should also be a View Source button, which when clicked would open the source for the current message in a new window. This would make it easy to check the full headers or otherwise see what an email has hidden inside it. (Currently, this functionality is available only as a global toggle which is dumb. You have to choose normal view, full headers view, or source view, and then after you are done looking at one message you have to go back and choose your default view again.)
The message display window toolbar needs a button that toggles the displayed message between a fixed font and a proportional font. By default, I want to look at my messages in a proportional font like Times Roman. But for those occasional messages with ASCII art, it would be nice to be able to flip it into a Courier font. Note that when the user prints the message, it should respect the current selection for proportional or fixed font. For extra credit, the setting should be remembered: if you get a message and click it over to fixed font, it should stay in fixed font forever until you change it back.
The message search feature is cool, but it could be better: by default, it should search both the Subject: header line and the From: header line (or, for the Sent folder, the To: line). If it searched both, I would almost never have to click on the selector that controls what it searches. Right now, I usually have to click on that. To all the people working on Evolution: It's great! Please keep up the good work.
Lindows is based on Debian. As far as I know, you should have no problem taking a Lindows notebook, removing all the packages that contain the strings "lindows" or "xandros", editing sources.list to point to your favorite Debian mirror, and doing apt-get dist-upgrade.
I am not a materials scientist, but here is what my expert bicyclist friend told me:
Aluminum is much more at risk for metal fatigue than steel or titanium. Because of this, the manufacturer does not dare design an aluminum frame that flexes, at all. Thus alumnium frames have to be designed to soak up shock without flexing (perhaps by adding carbon fiber stays and fork).
Steel and titanium frames can be designed to allow flexing to soak up road vibration.
Titanium can flex just about forever without fatiguing; it's just all-around strong and flexible. Its only real negative against steel is the cost.
You say that frames don't flex significantly, but in my experience different frames feel different when you ride them.
...low weight in rotating components is even more important. To accelerate a wheel or pedal and shoe system, kinetic energy of rotation must be supplied, in addition to the kinetic energy of linear motion. For example, with a wheel, if the weight is mostly concentrated in the rim and tire it would take nearly double the energy needed to accelerate it than an equal non-rotating weight. In other words, one pound added to a wheel or shoe/pedal system is equivalent to nearly two pounds on the bicycle frame.
A few years ago I completed a study with Ned Frederick of Exerter Research that showed how the addition of 250 grams to a shoe/pedal system or 500 grams total for both right and left feet (about the difference between the lightest and heaviest shoe/pedal systems on the market) would require an additional power output of 1.3 percent on rotational power requirements at constant speeds on a level road.
The effect on climbing, and accelerating, would be about 0.4 percent, but only during periods of climbing and accelerating. The effect on rolling resistance would be about 0.07 percent. This means that the net effect due to all factors of a realistic increase of 250 grams in mass on each leg would be to increase the power required by a cyclist by approximately 1.5 to 1.7 percent.
This increased cost could be significant in a long road race of 100 to 125 miles, with several hills and ending with a sprint. In terms of oxygen cost we estimated 1.3 % increase at 200 watts of work (approximately 24-25 km/hr) when only an additional 125 grams is added to each pedal/shoe system (total 250 grams for both legs).
But never mind the theory; we have observed that she rides faster on the new bike, especially uphill.
My wife and I are serious bicyclists. This year, we are attempting longer and harder rides than ever before, with more elevation gain. So we got new bikes.
My wife's bike is built on a frame by Wylder, called the Queen of the Road. Wylder is going out of business; when they were in business, they charged $1800 for that frame, but they are selling the frames they have left for $750. (Their web site says they have 42 cm, 45 cm, 52 cm, and 55 cm frames left. If you know a woman who wants a ti bike and can fit one of those sizes, send her this link!)
My bike -- I'm still amazed, I feel so lucky -- is a Colnago Ovalmaster. It's 6/4 titanium, the extra-stiff kind, and it has oddly shaped tubes; the shaped tubes help make the bike very stiff in the ways I want it to be stiff (when I hammer on the pedals, the frame doesn't flex) but still light and springy when I go over bumps. I love it.
I was able to afford my frame because I got it used on ebay. You can get some great deals on bike stuff on ebay.
Anyway, our new bikes have really helped us improve our speed. I didn't realize how much my wife's old bike was holding her back, until she got the new one. Her old bike is comfortable and stable, but it's six full pounds heavier than her new one, which matters when you are spending many hours climbing tall hills. And a lot of the weight savings is in moving parts: pedals, cranks, wheels. (The rule of thumb is that rotating weight counts double, so you should sooner get a lighter wheel than a lighter saddle.)
For me, the biggest improvement with the new bike is not the climbing, but the descending! My old bike was not stiff enough for me, and I got very nervous going down a steep hill. My new bike feels rock-stable under me when I am descending. I am still not a speed demon on downhills, but I'm a lot happier and a bit faster.
The metric I like to apply to bikes is dollars per mile. We ride enough miles that even just this year, both bikes will drop to well under a dollar per mile. If you only ride 10 miles at a time, don't spend serious money on a bike; get an inexpensive starter bike. (But don't buy a $200 bike from Wal-Mart. If you are considering buying a new bike, please read my thoughts on my personal web page about bikes.)
When you think about it, there is a lot of technology going with us on bike rides. We have heart rate monitors that also record speed, distance, and altitude changes (Polar S720+). We have technical fabric clothing, such as CoolMax jerseys or microfiber polyester rainwear. We have helmets that can save us in a bad accident, yet are light and don't make our heads overheat. We carry cell phones while riding, in case a bike totally breaks down and we need a taxi, or in case someone needs medical help.
No matter how much technology I buy, the pro riders could still ride me into the ground on a low-end bike. But our titanium bikes, and our other gear, let us ride up to our ability and have fun doing it.
You hit the back button and you expect it to go to the previous step in the process. Sometimes it does this, sometimes it gives you a page expired.
In other words, you want the behavior of the back button to be slightly improved. Andreeson is complaining because we still have a back button.
Now suppose you finish the checkout process and you hit back. You wouldn't want to check out again, you'd want it to jump to whatever you were doing before you began the checkout process.
Actually, no, I wouldn't. I like my back button to be predictable. I want it to take me back one page. The scenario above is why I like right-click on the back button to drop down a history list; I'll skip the checkout by picking from the list.
I dont agree with your disk-space argument however.
Not sure what you mean by this. I stated a simple fact: back in the day, it cost 32KB per bookmark. And back in the day, disks were small.
These days, people are using better file systems (even FAT32 is better than FAT16) so it will cost 4KB, maybe less, per bookmark. If you did a scheme like this on ReiserFS, which can store very small files very efficiently, the overhead is lower. But never mind all that, because these days you can buy 180GB disks.
We have more or less roughed out what we all want from a browser. We like the back and forward buttons, etc. We are comfortable with them.
I suppose he's shocked that after decades of research, cars still come with a steering wheel and a gas pedal, instead of something futuristic.
Now, we not only have things like tabbed browsing, but we have more subtle things that are still nice. For example, in Galeon (for Linux, at least) you can click on the New Tab button with the middle mouse button instead of the primary one, and it will open a new tab with the URL from the selection buffer. So now, instead of:
0) Select URL 1) Click New Tab button 2) erase URL in location bar (be careful not to select it!) 3) click middle mouse button in location bar 4) hit Enter key to load URL
you can just do:
0) Select URL 1) click middle mouse button on New Tab button
It's not earth-shaking, but I like it.
Now take that one feature, and all the other little tiny nice features, and roll them all up. It may be subtle, but it's progress and I'm happy.
Well IE is sort of better at this, in that favorites are individual files, so you can use the filesystem's find function to search (nice when you have 1000+ bookmarks).
Oh, I hate that one-file-per-bookmark idea. You aren't allowed to call the bookmark whatever you want -- why did they disallow characters like '?' or ':', instead of BASE64-encoding them or something? And these days it's not so bad, because most people are running FAT32 or something better, but back in the day there were a lot of people running FAT16, and on a 2GB disk partition, each bookmark used up 32KB of storage! Yikes!
I'd rather just have a non-sucky UI for finding inside the bookmarks file. (I've just started using Mozilla Firebird and so far the bookmarks searching seems pretty good.)
Slashdot Mirror
The fact that the language under critque was FORTRAN, unfortunately today, obscures the underlying truths they were discussing.
I disagree. That book rocks; every software geek should read it. Yes, some of the examples are non-interesting because they are too FORTRAN-related. But lots of the other examples are timeless classics. The discussion of floating-point math is a gem. "Floating point math is like moving piles of sand around; every time you do it you lose a little sand and pick up a little dirt." (Quote from memory so it might not be exact)
And by the way, about half the examples are in PL/1, not FORTRAN.
steveha
One of my favorite features in Evolution is the safe preview: when you look at an email message, it renders the message in all ways that do not involve hitting a server. So an HTML message with bold, italics, colors, pictures, etc. will display correctly... except for anything that would touch a server.
Why is this important? Because spammers make special URLs that encode your email address, and their servers use the special URLs to track you. If you even preview the message in Outlook, or in Mozilla Mail, their server gets a hit and they know your email address is a good one.
There is a command on a menu, "Load Images", that will go ahead and put hits on servers and render the message completely. You can use this for email from sources you trust. (It ought to be a toolbar button, but it isn't yet in Evolution.)
If Thunderbird doesn't already have this, I ought to file a bug.
steveha
You'll see from the email archives that they are busy implementing the latest JDK specifications, which are publically available.
2 759.html
Okay, I checked the email archives, and I found this:
http://www.kaffe.org/pipermail/kaffe/2003-June/04
You said that "All of the Java platform is available on a free license for open-source developments", but it looks to me like Sun is locking down anything new under the SCSL. SCSL isn't a free license; the Kaffe guys say so, and I agree with them.
So, are you saying that the SCSL is a free license, or are you saying that Sun is releasing new Java stuff under some other license, or what?
Can you give me a link to some web page or email or something that says that I'm wrong and Sun released the latest Java specs under a free license?
steveha
You are mistaken about Microsoft and OS/2. Microsoft really believed that OS/2 was the future. I worked at Microsoft in the early 1990's and everything was all OS/2, OS/2, OS/2; developers all had OS/2 computers for development work, the computers in the library ran OS/2, all Microsoft applications had OS/2 versions available, etc.
Customers voted with their dollars, and they voted for Windows rather than OS/2. I believe this was due mainly to the fact that Windows had a much easier migration path: if you had several DOS apps that you needed, you could run them all in Windows, versus running one at a time in the compatibility box under OS/2 and possibly crashing your computer. (Yes, later OS/2 versions were better, but that was after Windows had already won and Microsoft was already gone.) Other issues were that Windows ran much better on the computers that people had back then, and that Windows cost less than OS/2.
So, once Microsoft figured out that the customers wanted Windows and didn't want OS/2, Microsoft made the famous deal with IBM where IBM got OS/2 and Microsoft kept Windows. Microsoft didn't betray any OS/2 users, because IBM was there to support those OS/2 users.
In summary, Microsoft didn't have some cynical bait-and-switch plan, because internally Microsoft was pushing OS/2 right up until the famous "divorce" from IBM. And Microsoft didn't "pull the rug out" because IBM was fully supporting OS/2. It's not Microsoft's fault if IBM wasn't able to take over the world with OS/2.
Microsoft does have some things to answer for, but this really isn't one of them.
steveha
FACT: All of the Java platform is available on a free license for open-source developments, including the test suites. This is what the Kaffe people use. Nothing comparable exists for Dotnet whatsoever.
I wish things were as rosy as you make them sound. Kaffe is stuck in Java 1.x compatibility, because Sun Microsystems is keeping the Java 2.x specifications unfree. See section 2.3.1.5 , "Why is (some) free software not implementing Java2?", of the Debian Java FAQ.
steveha
...and I do it under Linux.
I have a classic, wonderful monitor: an SGI 1600sw, with 1600x1024 resolution. I only run it in its native resolution. My fonts are large and beautiful.
0) Use TrueType fonts.
Make sure your X11 setup is all correct for scaled fonts, especially TrueType fonts, and then get some good ones. Go everywhere and make sure you are using your good TrueType fonts. My GNOME preference fonts are all TrueType, plus my web browsers. The GNOME 2.x dialog for this is Applications / Desktop Preferences / Font. If you have an LCD flat panel display, be sure to check the box that says "Subpixel smoothing (LCDs)". For a CRT monitor, I suggest you check the "Best shapes" box.
1) Grow your fonts.
Go into XF86Config (or, in Debian, XF86Config-4 if you are using a recent version of XFree86). Find the part where it describes the monitor. There should be a DisplaySize line describing how big your monitor is, in millimeters. If the line is not there, search the web for specs on your monitor, or just measure it, and add the DisplaySize line. For the 1600sw:
DisplaySize 369.4 236.4
Now we want to lie to X11 about the size of our monitor, and say it's smaller than it really is. I want fonts 150% the usual size, so I multiply each number by 100/150 (i.e. 2/3 or 0.66666).
DisplaySize 246.3 157.6 # lie to get 150% font zoom (165 dpi)
# DisplaySize 369.4 236.4 # correct: 1600x1024 at 110dpi
Note that I like to leave comments about what the heck I'm doing here and why.
Now, X11 thinks my monitor is 165 DPI, instead of the real 110 DPI. When an application asks X11 to display 12 point text, X11 scales the TrueType font accordingly. I get automatic, across-the-board font zoom.
Peeve: there ought to be an X11 setting for this. You ought to be able to specify a zoom level, say 150%, and have X11 honor it without bastardizing the monitor size. If I can't have a zoom level setting, at least let us specify the DPI as a DPI number, instead of as the number of millimeters our monitors are!
2) Grow your web fonts too
Now your other big problem will be web sites that hard-code sizes. Even with 150% zoom, you really don't want 6-point fonts. The "minimum font size" setting in Mozilla hasn't worked well for me when I tried it in the past. You can specify a horrid large font size in the prefs, but then when you print a page, it prints huge too!
The solution is to use a cascading style sheet (.css) file. Go to your ~/.mozilla/default/<something>/chrome directory, and edit a file called userContent.css. (Be sure to check out the example files that Mozilla leaves there for you, while you are there!)
Add these lines to userContent.css and save:
@media screen {
* {
font-size: 28px !important;
line-height: 30px !important;
}
}
These lines mean: only for display on the screen (not while printing!), set the font size to 28 pixels height, and the line height to 30 pixels height. The "!important" part means you insist, even if the web page specifies a smaller size.
Now revel in the easier-to-read text.
You still have problems. Web designers who lay out pages with tiny fonts didn't expect their fonts to be forced huge, so the page won't look right; it might look downright ugly. And this fix does nothing to help when the webmaster specified a column width in pixels, so you may find a column that was intended to be over half your screen width is actually only three inches wide! (Thus you have big, easy-to-read text in a skinny very tall column, and you have to scroll the page a lot to read it.)
You also may find some text-entry forms that are 6 points tall, but the text you are typing into them is still huge, so you can't really read it. I ought to figure out what preference sets minimum text-entry box size.
Anyone with more useful tips, please share them!
P.S. Slashdot would not let me include the lines from my config files
I have a Tungsten T and I love it. When the newer Palm devices came out, I said to myself, "I wish Palm would come out with a device just like the Tungsten T but with the new tranflective screen". Here's the T2, and it fits the bill.
Double the memory is nice, the transflective screen is great, and the rest is almost purely unchanged. That spells a winner in my book.
I'm nervous about Graffiti 2; I like Graffiti as it is. But Palm lost a lawsuit over Graffiti and they don't dare ship it in new models, so I don't blame them for the change, and perhaps Graffiti 2 will be a good thing anyway.
It also sounds like they have sweetened the software bundle, also a good thing.
Some posters have commented that this is "nothing new". I say it's exactly what I wanted. A Tungsten T is a sweet little PDA, and the T2 should be better still.
steveha
Blazer uses a proxy server. Handspring says so.
Your wider point, that there are choices in the Palm market, is spot-on. If Palm were to die and take their proxy with them, you would be able to get some other browser for your palm.
steveha
...I don't see any positive benefit to purchasing the license unless they intend to go after each business individually in court.
However, they are hinting that they will do exactly that. They sent letters to 1500 companies they believed were running Linux, saying that the companies themselves might be liable (i.e. that SCO might sue them later).
In practice, it would be a very big job to go after a bunch of companies, each in turn, one at a time. But that's the threat they are using to encourage companies to buy their licenses.
steveha
I'll second the idea of bicycling to work. I was working far too many hours at a desk, and I started riding a bike to work; it made a difference for me.
It took me 20 minutes to drive to that job, or about 50 minutes to ride my bike to that job. So biking really only cost me a net 30 minutes each way. I spent 60 minutes to get 100 minutes of exercise; such a deal.
I was riding 10 miles each way to work. If you live too far away from work, do a bus/bike or car/bike combination: ride partway on the bus or car, then go the rest of the way on your bike. (I'm assuming that the buses near you have bike racks on the front, like the ones near me. If not, maybe you can arrange to store a bike somewhere that the bus can take you to.)
The secret of happy bike commuting is in the equipment. You need good clothing, including rain gear (jacket, helmet cover, shoe covers, etc.). You also need a good bike, with fenders and a cargo rack.
Stash some clothes at work, so you don't have to haul lots of clothing every day when you ride.
Bicycling is a good aerobic exercise, great for burning calories. It will work even better if you can combine it with weight lifting or some other exercise that works your arms, back, etc. and not just your legs and cardiovascular system.
steveha
Red Hat could box up a single disk. This disk would install a Red Hat base system, which would then be able to use the Internet to download the rest using Up2Date.
Note that this one disk, if done correctly, could be distributed pretty much forever. Currently, when a new version of Red Hat comes out, a bunch of the old version goes into the trash. The one-disk installer wouldn't have that problem.
Note also that Debian works this way now too. The PGI installer is one CD (and that one only about 1/6th full) but it installs the latest and greatest versions of all your packages, from the Internet. There is even a bootable Linux recovery disk that can, as one of its features, install a Debian base system that you could bootstrap into a full Debian system.
Anyway, a single installer CD in a shrinkwrap box is all they need, as long as there is info on where to order a set of the latest packages on CDs for a CD install. You should also be able to buy those CD sets from CheapBytes or similar distributors.
steveha
But remember the "Freedom Zero" debate. Tim O'Reilly commented that the first freedom of free software, and he called it Freedom Zero, is the freedom to choose which license you use to release it. RMS rejected this concept, calling it "Powerplay Zero", saying in effect that only the GPL is acceptable.
ESR then challenged RMS publicly: if you had the power to require all software to be GPL software, would you do it? RMS did not answer. A "yes" answer would be consistent with his other public statements, however.
So I don't know if we can prove that one of RMS's goals is the eradication of commercial software, but his public statements are consistent with that position. A relevant quote from the "Powerplay Zero" essay:
Proprietary software is an exercise of power -- it harms the users by denying their freedom.
By the way, RMS has never been opposed to people charging money for software; he just insists that the software must be under the GPL (and thus come with source code, be subject to forks, etc.). This, in the real world, means that there is a (fairly small) upper bound on how much one can charge for the software.
steveha
[IBM] can hold those patents (and charge others for the use of them for non-GPL'd applications) and GPL code based on those patents.
Yes, exactly correct. This is why Linux has RCU and BSD doesn't. IBM is willing to license the patents for free for GPL code, but they still want to charge license fees for use in proprietary code.
Releasing something under BSD is effectively placing it in the public domain.
By the way, you sometimes see people claiming that "if there weren't any copyrights there would be no need for GPL" or some such. Not so. GPL uses copyright law to prevent people from taking free projects proprietary; BSD lets you do anything you want, including taking a free project, hacking it up slightly, and releasing it as a proprietary product.
steveha
It would be a hefty task for the filesystem to transparently convert /album into the appropriate ID3 (or whatever it is for ogg) format while allowing this functionality for every conceivable format (mp3, ogg, wmf, ra, ...).
Plugins. Customizable plugins that run in user space. The article talked about this a little bit.
steveha
Just on general principle, I prefer my data inside the file and not left with the filesystem. The MP3 metadata example, to me, is like Windows file extensions on HGH.
.attributes created on the target legacy system; I'd be happier if just one big XML file could be created with the same name as the original file.
//rich onto server //legacy, and then you want to restore some files from //legacy to //rich. If all the metadata was stored in a big XML file, then when you copy the file from //legacy to //rich you restore all the metadata; you wouldn't accidentally slice off attributes by forgetting to copy one or more rich attributes files.
I'm with you -- I like self-contained file formats.
But I don't think he was proposing that you not use Ogg tags or MP3 tags; he was talking about the filesystem abstracting the tags. If you changed "Stagnation.ogg/album" to the string "Trespass", then the filesystem abstraction layer should update the Ogg "album" tag inside the file to be "Trespass".
The key benefit here is that you would not need some wacky command-line utility program to let you view and change tags on Ogg files. You could just use the shell. In bash:
for ii in *.ogg; do echo "Trespass" > $ii/album; done
Note that this same one-liner would work if you were in a directory with MP3 files, and you changed "*.ogg" to "*.mp3". Currently, you need to run vorbiscomment for your Ogg files, and mp3info for MP3 files. (I just checked, and sure enough, they take different arguments to do the same operations.)
Personally, I'd like to see a standard metadata portable XML format for legacy systems. People talk about copying a file from a rich metadata filesystem and having new files like
Suppose you backup server
You could do most of the fancy tricks of the rich metadata filesystem on a legacy filesystem that used the big XML file to store the rich metadata. And as long as the legacy system is just smart enough to look at the main data part of the XML and leave the metadata tags alone, you could still modify the file with sed, awk, perl or whatever, and then copy the big XML file onto your rich metadata filesystem and still not lose any rich metadata.
Note also that the big XML file could be used to deal with existing rich metadata systems, like resource forks from Macintosh filesystems, or multiple data streams from NTFS files.
steveha
Next to the delete button, a "spam" button. This would delete the mail message, but also flag the message as a spam. You should be able to customize this to run your local spam filter training script.
In the message display toolbar, there should be a button that loads images. By default, Evolution does not render any HTML that puts a hit on a server; this is a great feature, because spammers do use tricky URLs that encode your email address. When their server gets a hit from the URL with your address encoded, they know your email address is good, and you get more spam. Currently, the "Load Images" command is in a menu, and is not available with one click. "Load images" should also be in the context menu (the "right-click" menu). For extra credit, Evolution should remember that you already loaded the images, and if you go back again it should be able to get the images from the browser cache, not have to hit the server again as it currently does.
There should also be a View Source button, which when clicked would open the source for the current message in a new window. This would make it easy to check the full headers or otherwise see what an email has hidden inside it. (Currently, this functionality is available only as a global toggle which is dumb. You have to choose normal view, full headers view, or source view, and then after you are done looking at one message you have to go back and choose your default view again.)
The message display window toolbar needs a button that toggles the displayed message between a fixed font and a proportional font. By default, I want to look at my messages in a proportional font like Times Roman. But for those occasional messages with ASCII art, it would be nice to be able to flip it into a Courier font. Note that when the user prints the message, it should respect the current selection for proportional or fixed font. For extra credit, the setting should be remembered: if you get a message and click it over to fixed font, it should stay in fixed font forever until you change it back.
The message search feature is cool, but it could be better: by default, it should search both the Subject: header line and the From: header line (or, for the Sent folder, the To: line). If it searched both, I would almost never have to click on the selector that controls what it searches. Right now, I usually have to click on that.
To all the people working on Evolution: It's great! Please keep up the good work.
steveha
Lindows is based on Debian. As far as I know, you should have no problem taking a Lindows notebook, removing all the packages that contain the strings "lindows" or "xandros", editing sources.list to point to your favorite Debian mirror, and doing apt-get dist-upgrade.
steveha
I am not a materials scientist, but here is what my expert bicyclist friend told me:
Aluminum is much more at risk for metal fatigue than steel or titanium. Because of this, the manufacturer does not dare design an aluminum frame that flexes, at all. Thus alumnium frames have to be designed to soak up shock without flexing (perhaps by adding carbon fiber stays and fork).
Steel and titanium frames can be designed to allow flexing to soak up road vibration.
Titanium can flex just about forever without fatiguing; it's just all-around strong and flexible. Its only real negative against steel is the cost.
You say that frames don't flex significantly, but in my experience different frames feel different when you ride them.
steveha
Here is an explanation:
Dr. Ed Burke weighs in on lightweight mountain bike components
The most relevant excerpt:
But never mind the theory; we have observed that she rides faster on the new bike, especially uphill.
steveha
My wife and I are serious bicyclists. This year, we are attempting longer and harder rides than ever before, with more elevation gain. So we got new bikes.
My wife's bike is built on a frame by Wylder, called the Queen of the Road. Wylder is going out of business; when they were in business, they charged $1800 for that frame, but they are selling the frames they have left for $750. (Their web site says they have 42 cm, 45 cm, 52 cm, and 55 cm frames left. If you know a woman who wants a ti bike and can fit one of those sizes, send her this link!)
My bike -- I'm still amazed, I feel so lucky -- is a Colnago Ovalmaster. It's 6/4 titanium, the extra-stiff kind, and it has oddly shaped tubes; the shaped tubes help make the bike very stiff in the ways I want it to be stiff (when I hammer on the pedals, the frame doesn't flex) but still light and springy when I go over bumps. I love it.
I was able to afford my frame because I got it used on ebay. You can get some great deals on bike stuff on ebay.
Anyway, our new bikes have really helped us improve our speed. I didn't realize how much my wife's old bike was holding her back, until she got the new one. Her old bike is comfortable and stable, but it's six full pounds heavier than her new one, which matters when you are spending many hours climbing tall hills. And a lot of the weight savings is in moving parts: pedals, cranks, wheels. (The rule of thumb is that rotating weight counts double, so you should sooner get a lighter wheel than a lighter saddle.)
For me, the biggest improvement with the new bike is not the climbing, but the descending! My old bike was not stiff enough for me, and I got very nervous going down a steep hill. My new bike feels rock-stable under me when I am descending. I am still not a speed demon on downhills, but I'm a lot happier and a bit faster.
The metric I like to apply to bikes is dollars per mile. We ride enough miles that even just this year, both bikes will drop to well under a dollar per mile. If you only ride 10 miles at a time, don't spend serious money on a bike; get an inexpensive starter bike. (But don't buy a $200 bike from Wal-Mart. If you are considering buying a new bike, please read my thoughts on my personal web page about bikes.)
When you think about it, there is a lot of technology going with us on bike rides. We have heart rate monitors that also record speed, distance, and altitude changes (Polar S720+). We have technical fabric clothing, such as CoolMax jerseys or microfiber polyester rainwear. We have helmets that can save us in a bad accident, yet are light and don't make our heads overheat. We carry cell phones while riding, in case a bike totally breaks down and we need a taxi, or in case someone needs medical help.
No matter how much technology I buy, the pro riders could still ride me into the ground on a low-end bike. But our titanium bikes, and our other gear, let us ride up to our ability and have fun doing it.
steveha
You hit the back button and you expect it to go to the previous step in the process. Sometimes it does this, sometimes it gives you a page expired.
In other words, you want the behavior of the back button to be slightly improved. Andreeson is complaining because we still have a back button.
Now suppose you finish the checkout process and you hit back. You wouldn't want to check out again, you'd want it to jump to whatever you were doing before you began the checkout process.
Actually, no, I wouldn't. I like my back button to be predictable. I want it to take me back one page. The scenario above is why I like right-click on the back button to drop down a history list; I'll skip the checkout by picking from the list.
steveha
Um, Galeon has "middle click on link to open link in a new tab" too.
What does Mozilla have for selecting a URL from a text file or something, and opening it in a new tab?
steveha
I dont agree with your disk-space argument however.
Not sure what you mean by this. I stated a simple fact: back in the day, it cost 32KB per bookmark. And back in the day, disks were small.
These days, people are using better file systems (even FAT32 is better than FAT16) so it will cost 4KB, maybe less, per bookmark. If you did a scheme like this on ReiserFS, which can store very small files very efficiently, the overhead is lower. But never mind all that, because these days you can buy 180GB disks.
steveha
We have more or less roughed out what we all want from a browser. We like the back and forward buttons, etc. We are comfortable with them.
I suppose he's shocked that after decades of research, cars still come with a steering wheel and a gas pedal, instead of something futuristic.
Now, we not only have things like tabbed browsing, but we have more subtle things that are still nice. For example, in Galeon (for Linux, at least) you can click on the New Tab button with the middle mouse button instead of the primary one, and it will open a new tab with the URL from the selection buffer. So now, instead of:
0) Select URL
1) Click New Tab button
2) erase URL in location bar (be careful not to select it!)
3) click middle mouse button in location bar
4) hit Enter key to load URL
you can just do:
0) Select URL
1) click middle mouse button on New Tab button
It's not earth-shaking, but I like it.
Now take that one feature, and all the other little tiny nice features, and roll them all up. It may be subtle, but it's progress and I'm happy.
steveha
Well IE is sort of better at this, in that favorites are individual files, so you can use the filesystem's find function to search (nice when you have 1000+ bookmarks).
Oh, I hate that one-file-per-bookmark idea. You aren't allowed to call the bookmark whatever you want -- why did they disallow characters like '?' or ':', instead of BASE64-encoding them or something? And these days it's not so bad, because most people are running FAT32 or something better, but back in the day there were a lot of people running FAT16, and on a 2GB disk partition, each bookmark used up 32KB of storage! Yikes!
I'd rather just have a non-sucky UI for finding inside the bookmarks file. (I've just started using Mozilla Firebird and so far the bookmarks searching seems pretty good.)
steveha