Re:Why is LISP superior?
on
RMS The Coder
·
· Score: 1
I started learning Lisp when I started using EMACS, then for a while I was hacking in Scheme (the Programming Language Concepts class at my alma mater was taught by a fellow who received his degree from Indiana University), and now I am doing a lot of programming (in my Copious Free Time) in Common Lisp - and I must say that of the three, I like Common Lisp the best.
Out of curiosity, why do you like Common Lisp over Scheme? I haven't actually used Common Lisp, but I've used a few other Lisp variants including Scheme. Scheme seemed to be the "nicest" of all the ones I tried.
Oh yes, and how about we shoot pronographers, because pornography is wrong. And we can also lock up people who use cuss words, because that's wrong. Oh, and non-Christians can also be thrown in jail, because that's just plain wrong. And people who wear plaid. Plaid is wrong.
If free software has to resort to force, then it is no longer really free, now is it? The freedom of choice is a much more basic right than the freedom to duplicate.
Non-free software hurts no-one. If you don't like it, don't use it. What about software that isn't "open source" in the OSD sense, but does come with source code? (ie: you get the source when you buy it, but you're not allowed to distribute it) How does that hurt anyone? All it does is ensure that the developers actually get paid for the work they did.
Actually, it's interesting that you mention that. All of the people I know personally who have wrist trouble are emacs users. I'm a VIM user myself, and I've never had wrist problems.
Chording is supposed to be pretty hard on the wrists. VI and its derivatives don't use chords for very many things, except for a few shifts, and a very small number of rarely used control keys.
A possible alternative to a vi-like editor would be to use emacs with "sticky keys". Knowing emacs, there's probably an elisp script that can turn that on for you. Then you's press (and release) each meta key before the key it modifies. So "C-x" would be 'control' followed by 'x'.
Or you could use one of the vi-modes for emacs...:-)
Pentium is the same as ununpentium (un's are basically not's, so they cancel). Ununpentium is element 115. Proof that Intel has technology reverse engineered from alien toaster ovens!
Yeah, every software company in the early eighties was either "Micro-*" or "*-Soft". "Micro Soft" was about the most generic software company name in existence.
About words sounding okay after hearing them a lot, that reminds me of when I first heard that movies with sound were originally called "talkies". It sounds absurd. Then I started thinking about the word "movie". It's just as dumb. Pictures that move... they're kind of move-ee, they're movies... ugh.
Actually, it was "Commodore VIC-20". VIC is an acronym for Video Interface Chip, and "Commadore" isn't a real word.
Incidently, the 20 part is sort of funny. The whole machine was designed around the VIC chip, but they wanted to add a number to it. 22 was the obvious number, because the display was 22x23 characters. "VIC-22" doesn't sound very nice though (22 sounds like a gun caliber...), so they changed it to 20, which is "friendly sounding". So the 20 doesn't really quantify anything, except maybe a very rough approximation of the number of columns...
That reminds me, I have to get my VIC-20 back from that guy...
What possible use can a cookie provide except for exploitation of those that visit your site? For any example you provide I can provide an alternative method for saving info on the web host's drive (secured and with the client's permission of course) that would accomplish the same task.
And then every time people visit your site they have to log in with a userid and password, or a special PIN. What about moving from one page to the next (ie: for a shopping cart)? You could store the state in the URL I suppose. But none of those really buy you any extra security. Infact, they might buy you less. I've seen systems where the userid and password were stored as clear-text in the URL's.
The thing to remember with cookies is that the information comes from the server. In other words, it's information they already have. All cookies give web designers is persistent state. For most sites that persistence only lasts for the current session as well. Cookies are only a problem for the ultra-paranoid.
Those who benefit from the current system where programs are property offer two arguments in support of their claims to own programs: the emotional argument and the economic argument.
The emotional argument goes like this: ``I put my sweat, my heart, my soul into this program. It comes from me, it's mine!'' This argument does not require serious refutation. The feeling of attachment is one that programmers can cultivate when it suits them; it is not inevitable. Consider, for example, how willingly the same programmers usually sign over all rights to a large corporation for a salary; the emotional attachment mysteriously vanishes. By contrast, consider the great artists and artisans of medieval times, who didn't even sign their names to their work. To them, the name of the artist was not important. What mattered was that the work was done--and the purpose it would serve. This view prevailed for hundreds of years.
According to RMS, those Apple employees should just be happy that the work was done, and shouldn't be concerned with having their name on it. After all, getting credit for something is like saying you own it, and we all know* that owning software is evil.
Incidently, I do agree with you that the definition of "hacking" probably shouldn't include any moral elements. Hacking does imply something that is clever, however, and creating unmaintainable code is most decidedly un-clever.
In fact, that whole topic of moral sanguinity makes a huge and unnecessary gray area. For instance, say you're a hired programmer and you write a solution to a problem which is backwards, nearly illegible on first glance, and absolutely brilliant (and, of course, functional). However, many would argue that you have some kind of moral obligation to your employer to write maintainable code. Does this mean what you did wasn't hacking but PVAC?
There already seems to be a gray area here... Persoanlly, I think that if your code is completely non-maintainable and illegible, then it's automatically disqualified from being "brilliant", and hence is sounds more like a kludge than a hack to me. Maintenance is 98% of the typical long-lived software project.
Re:Backspace considered harmful
on
Interface Zen
·
· Score: 1
You complain about the size of the backspace key. Why are you using a backspace key? ^H is much easier to get to, and it's always in the same place. Backspace is a useless key. Just xmodmap it awayand put backtick and tilde there.
Because chording is slower. While I'm hitting backspace, my left hand can be getting ready to hit the next key. If I use ^H I either have to use both hands, hence slowing me down, or I have to use one hand to chord, which is painful and awkward (especially for a key like H which is almost as far away from both control keys as possible).
BTW: If you don't believe that my left hand can be getting ready to hit the next key, try typing "stargaze" and "shamrock". Most people can type the latter significantly faster, because it requires alternate hands when typed on a qwerty keyboard.
I think you mean "flow", and other comments...
on
Interface Zen
·
· Score: 1
The "zenning out" you refer to is usually called "flow" I believe. It's discussed in Alan Cooper's "About Face" on page 127. Cooper is rather adamant about creating UI's that don't disrupt the flow. He's very much against useless modal dialogs, for example.
I'm a VI user myself (actually, I usually use VIM). I don't have any problem with where escape and control are on PC keyboards. The escape key is farther, but not much farther. It's on the corner though, which I find makes it much easier to find by feel. When I go to press escape, I just reach for the corner key with all four (non-thumb) fingers, without fear of hitting anything else. The key isn't bigger, but by using several fingers it balances out.
As for control, it's also in the corner on PC keyboards, which I find makes it relatively easy to find by feel. I do agree that caps-lock in in an exceedingly dumb place though. Having it be a littel "chiklet" up by the caps-lock light would probably be more appropriate.
Regarding your complaint that there should be control, alt and "vanity" keys on the right side of the keyboard too... huh? Every PC keyboard I've seen *does* have the shifting keys on both sides of the keyboard. That's part of the reason the space bar is smaller than it used to be. You can't have it both ways.
Finally you had several complaints about the cursor-key layout. I think the upside-down T is pretty good, actually. Moving my middle finger between the up and down keys is very easy, and I can keep fingers on the left and right keys in either position. I rarely want to go both up and down simultaneously...
Again, by being segregated, the cursor keys are fairly easy to find by feel, and the distinctive shape makes it easy to position your fingers on the keys without looking away from the screen. It's much better than the cluster of six keys above them.
Sun keyboards also have annoyingly small backspace keys. Maybe you never make a typo, but I use backspace quite frequently...
Finally, as for mice, yes there is a big lag involved in switching between the keyboard and the mouse. Well designed software takes that into account, and has keybindings wherever possible. Unfortunately, many X apps are worse about this that their MS-Windows counterparts.
Ideally, software should be designed such that if you *have* to use the mouse (ie: in a paint or drawing program), you should be able to keep your hand there for a long time, and have one-key accelerators for your non-mouse hand to use. Deluxe Paint III on the Amiga did this. Actually, Star Craft does this as well...
My big keyboard nit (this is actually a software one): why the #&!! are the default behaviour of the backspace and delete keys on Linux (at least, on all the distributions I've used) so screwed up? Backspace should be ^H, and delete should be ^?. Yes, I know how to fix it, but the default is dumb, IMHO.
Java applets already have a restriction preventing them from opening connections to any machine except the one they came from. To get around this applets must be signed, and you would have to accept the corresponding certificate.
As others have pointed out, adding such restrictions to images and frames would break too many things. If you really want to, you could try adding such a feature to Mozilla, and see what it does.
In any case, I believe this Comet Cursors thing uses a plugin. If that's the case, it could probably just go and open a socket connection on its own. The same goes for any other plugin or Active X control (which is what I assume you meant by "embedded controls").
Postscript isn't going away any time soon, and the glitches you mentioned will be eradicated over time in classic open-source style. The turing-complete problem isn't a horrible problem because the postscript-generating program can limit itself to generating postscript commands that are known to produce predictable results.
I'm not suggesting that PostScript go away. I don't mind printers understanding it. I just don't think it should be the API that application programs use to talk to the printer.
And no, nothing, not even the open source movement, will erradicate the halting problem. It's been proven mathematically.
Finally, saying that an application program can "limit itself to generating postscript commands that are known to produce predictable results" is similar to saying OS's shouldn't have memory protection because applications can "limit themselves to accessing memory in a ways that are known to produce predictable results". The fact is, many applications don't. Many times I've tried to print things "2up", only to find that the application was generating postscript that circumvented it. A proper printing API wouldn't allow that.
And even if it weren't for these problems, does it really make sense that application writers should need to learn PostScript in order to print? That's like requiring people to learn the X protocol if they want a GUI. There should be an actual API something like Xlib, or at the very least a simpler page description language, like one based on XML.
There is no technicial problem in the following code, style is another issue, but compile it with 'gcc -pedantic -Wall' and gcc will complain "`x' might be used uninitialized in this function."
Yes, but if you look at the code in ghostscript, you'll see that it isn't a trivial situation like that. Most of the places I looked at had switch statements inside loops. I haven't investigated it in great detail, but it looked suspiciously like it was scanning though the input, and initializing certain variables in only some of the cases in the switch statement. If that's correct, then it's possible that invalid input would make ghostscript use the values of uninitialized variables.
Even if the code was correct at one point, I doubt that it is now. Bad style generally leads to errors, especially in programs that are old and have many maintainers.
...the Linux version of WP 8 was pretty much a port of the DOS/Win product that did not take advantage of the Linux platform. I really thought it was quite silly to have application-specific printer drivers and fonts, in this day and age. That's so... 80s.
Don't get me wrong, I'll take Linux over Windows any day, but I'd have to say that one of Linux's weaknesses is no real support for system-wide printer drivers. Most Linux and Unix apps assume that you've got a PostScript printer. If you don't you have to use something like ghostscript to act as a filter for your printer.
There are numerous problems with this setup. For one thing, ghostscript is complicated, rather badly written in places, and crashes far more often than I'd like. To get an idea of how hokey the ghostscript code is, try compiling it with '-pedantic -Wall'. You'll see tonnes of "variable x may have not been initialized" warnings...
Using PostScript as the API for communicating printers is just a bad idea. PostScript is Turing complete. That means there's all sorts of analysis you simply can't do to the stream sent to the printer. It isn't even guaranteed that the printer will ever finish (the halting problem). It would be a lot simpler if there was a drawing API, sort of like a subset of X, but for paper. Or heck, an XML-based page description language. Just make sure it isn't Turing complete.
If Linux actually had decent printing system, then WordPerfect wouldn't need its own. As it is, printing is one of Linux's biggest weaknesses, IMHO.
Having lived in both San Francisco and Toronto, I'd have to say your comment is incorrect. Canadian culture is quite different from the US. Those differences just aren't as easy to see without actually being immersed in both cultures, because the media in both places are very similar. And I'd have to say Toronto is more diverse than SF. Canadian culture as a whole is certainly more tollerant of diversity than American culture.
I once read about a study that was done in towns that lie on the border between the US and Canada. The study was essentially an opinion survey. They then checked to see if there was any significant difference between the Americans and the Canadians. A significant difference was found.
Perl is probably the most widely used language for CGI because of the lemming effect. If it had to do with effeciency of the resulting code, perhaps a compiled language like C or C++ would have been used. (though as others have pointed out, interpreted languages like Perl and Python will be just as fast as C or C++ for things that spend most of their time doing string processing) If rapid development was the goal, then there are other languages (like Python) that are arguably easier to write in than Perl. Perl is the most popular now because it was the most popular before.
Perl was the first language to become popular for CGI. Python didn't exist at the time CGI first appeared (AFAIK), and the web was young. Enormous commercial websites didn't exist, so efficiency wasn't a huge issue. CGI writers were mostly just trying to do "cool stuff", not develop large maintainable systems, so ugly but relatively easy to write Perl code was perfect.
Because virtually all of the early CGI scripts were written in Perl, Perl became the de facto CGI scripting language. Now most CGI programs are written in Perl simply because that's what everyone else does. If you go to the bookstore and look for books on CGI, most will be the Perl variety. Do a search on Google for "cgi script" and the first link has "Free Perl CGI Scripts" in the description. There are even some servers where they'll only let you have CGI scripts if they're written in Perl. People think "Perl" when they hear CGI.
This is a vicious cycle of course. The more popular Perl is for CGI, the more first-time CGI writers assume you need to use Perl, so the more popular it becomes. Perl is to CGI as MS Windows is to computers. Newbies think that's what they need to use, because of sheer numbers. The typical CGI newbie doesn't know you can use anything other than Perl, in the same way that the typical PC newbie doesn't know that their PC can run an OS other than Windows.
For the record, my first CGI program was written in Perl. I was going to write it in Rexx, but the server I was going to use only allowed Perl. All of the CGI programs I've written since have been written in Python (and they typically run on machines where I am root).
An idea I heard about a while back was to prevent MS from buying anything to produce their software except labor, and hardware. They wouldn't be allowed to buy other companies (or divisions of other companies), nor could they make "deals". They also wouldn't be allowed to "distribute" software written by others. Essentially they would be forced to innovate, rather than buying the innovations of others.
MS is pretty sneaky, so care would have to be taken to make sure they don't find a loophole. The best way to do this is probably to say MS may not spend any money except on the things in a carefuly chosen list (which would include employee salaries, hardware, etc.).
MS could then no longer buy out all of the competition, nor could they act as simple distributors for other companies products. They'd actually have to create things on their own. It would be hard for MS to make this look like "punishment" from a PR point of view. After all, they say they want "the freedom to innovate", and this wouldn't prevent them from innovating. If MS really was as innovative as they claim, such a sentence wouldn't hurt them significantly. I'm willing to bet it'll damage them quite seriously however.
Build a robot with your old PDA as its brain. Most PDA's have a reasonably powerful CPU, with a very low power consumption - perfect for a little robot. You might even be able to do this without opening the PDA if it's got a connector for a cradle.
One statment I found very... odd about thhe article was the following:
Contrary to popular wisdom that creativity thrives on unlimited freedom and unconventional thinking, Mazursky and his colleagues discovered that rules were essential to producing the ads the judges found creative.
Huh? Are there really people who believe that "creativity thrives on unlimited freedom"? Sorry, but it ain't so. I've always felt that the truly creative are those that are able to work within the rules, and still come up with something original and interesting.
There are many art forms that impose strict rules and structure including many forms of poetry (sonnets, haiku, etc.), most forms of music, dance and architecture. Without some sort of structure these become babbling, noise, convulsions and heaps of rubble. Structure is good. The truly creative can work within structure to create beuty. The non-creative wannabe's whine about being repressed.
I think the meta-moderation is going too far. Multiple moderators should be enough to balance things out. Giving people more moderation points would server to balance more. If I thought a moderation was unfair, and I had moderation points, I'd "counter-moderate" the post.
Dummy moderation controls is probably a bad idea. Only the obsessive will use them. The people who actually care about doing something useful simply won't use the controls if they think they'll probably just be ignored.
Also, I have to say that the "auto moderation" thing is very nasty. Just because someone posted a good comment or two, it doesn't mean all of their comments are good. Let each comment stand on its own. Likewise, there is the ongoing debate about whether AC posts deserve to start at 0 or 1.
How about this:
1. Increase moderation points, and don't make it quite so hard to become a moderator.
2. Remove the auto-moderation thing. All comments start at 1 (even AC posts).
3. Add a "killfile"-like feature where each user can artificially bump-up/down the scores of posts on a per-poster basis. Each person's "killfile" will only affect the scores of posts as they read them. So if you think AC posts deserve to get 1 less than everyone else, you'd put "Anonymous Coward -> -1" in the "killfile". You could even put this in everyone's killfile by default so the behaviour would be identical to what it is now (by default). If this is done, I think the score range should be chnged to -5 to 5 (instead of the current -1 to 5) and the "killfile adjustment" range should be -10 to 10 (so I can bump a -5 to a 5 or vice-versa if I think the poster is deserving of it).
3. From the main page, if the user has moderation privledges, add a "Moderate" link in addition to the "Read More" link. The "Read More" link will be for reading only, no moderation controls. The "Moderate" link will bring one to a special moderation page.
4. The moderation page would have moderation controls of course. It would also not show the scores, and sorting by score wouldn't be an option while moderating. And your killfile wouldn't have any effect in moderation mode either.
5. Finally, I don't think it should show who posted the comment. I really think it should be a meritocracy, and showing people's ID's to the moderators prevents that. I realize that everything having no ID could get confusing (especially in conversations), so I think for each person that posts a comment in a discussion area, there should be a new "fake ID" generated. This fake ID would be used for all of their posts on the moderation page. Fake ID's would not be persistent. In one discussion area my fake ID might be "Hoopy Cow" in another is might be "Green Frood". That way you could still follow conversations in a discussion area, but the identities of the posters would be disconnected from what they are in "Read Mode" and from all other discussion areas. Note also that Anonymous Coward would be treated as another user, so it wouldn't be obvious that AC posts were from AC's. Again, comments would be judged more on content rather than who posted them. Oh, and almost forgot. Signatures wouldn't be shown in moderation mode of course.
I realize that someone could go and open both "read" and "moderate" to find out who made various posts. It would be possible to disable moderation of discussion areas the user has alredy read, but then they could just read as AC, and moderate as themselves. I think that the above suggestions would at least reduce the chance of people moderating based on identity rather than content though. Having the moderation controls shown right there with the poster's identity just makes it too easy to say "hey, I know that guy" and +1, or "oh, I hate him" and -1.
Oh, and some gratuitous other suggestions:
- Make plain-old-text really plain text. - spell checking! - link checking! - better conversion of busted URL's ("relative" URL's should probably be patched up to simply add a leading "http://", as that seems to be the most common error) - what's with those "parent" links? I can never figure out which one is really the parent when I'm in flat mode.
Slashdot Mirror
I started learning Lisp when I started using EMACS, then for a while I was hacking in Scheme (the Programming Language Concepts class at my alma mater was taught by a fellow who received his degree from Indiana University), and now I am doing a lot of programming (in my Copious Free Time) in Common Lisp - and I must say that of the three, I like Common Lisp the best.
Out of curiosity, why do you like Common Lisp over Scheme? I haven't actually used Common Lisp, but I've used a few other Lisp variants including Scheme. Scheme seemed to be the "nicest" of all the ones I tried.
Oh yes, and how about we shoot pronographers, because pornography is wrong. And we can also lock up people who use cuss words, because that's wrong. Oh, and non-Christians can also be thrown in jail, because that's just plain wrong. And people who wear plaid. Plaid is wrong.
If free software has to resort to force, then it is no longer really free, now is it? The freedom of choice is a much more basic right than the freedom to duplicate.
Non-free software hurts no-one. If you don't like it, don't use it. What about software that isn't "open source" in the OSD sense, but does come with source code? (ie: you get the source when you buy it, but you're not allowed to distribute it) How does that hurt anyone? All it does is ensure that the developers actually get paid for the work they did.
Actually, it's interesting that you mention that. All of the people I know personally who have wrist trouble are emacs users. I'm a VIM user myself, and I've never had wrist problems.
:-)
Chording is supposed to be pretty hard on the wrists. VI and its derivatives don't use chords for very many things, except for a few shifts, and a very small number of rarely used control keys.
A possible alternative to a vi-like editor would be to use emacs with "sticky keys". Knowing emacs, there's probably an elisp script that can turn that on for you. Then you's press (and release) each meta key before the key it modifies. So "C-x" would be 'control' followed by 'x'.
Or you could use one of the vi-modes for emacs...
Pentium is the same as ununpentium (un's are basically not's, so they cancel). Ununpentium is element 115. Proof that Intel has technology reverse engineered from alien toaster ovens!
No, it would be "Hexium", not "Sexium". Like in the periodic table: Ununpentium (115), then Ununhexium (116).
Windows powered is sort of lame because, well... it's an adjective!
"Hey look, I got this new PDA."
"What OS does it run?"
"Windows Powered"
"TYPE MISMATCH ERROR: OS name must be a noun"
Yeah, every software company in the early eighties was either "Micro-*" or "*-Soft". "Micro Soft" was about the most generic software company name in existence.
About words sounding okay after hearing them a lot, that reminds me of when I first heard that movies with sound were originally called "talkies". It sounds absurd. Then I started thinking about the word "movie". It's just as dumb. Pictures that move... they're kind of move-ee, they're movies... ugh.
Actually, it was "Commodore VIC-20". VIC is an acronym for Video Interface Chip, and "Commadore" isn't a real word.
Incidently, the 20 part is sort of funny. The whole machine was designed around the VIC chip, but they wanted to add a number to it. 22 was the obvious number, because the display was 22x23 characters. "VIC-22" doesn't sound very nice though (22 sounds like a gun caliber...), so they changed it to 20, which is "friendly sounding". So the 20 doesn't really quantify anything, except maybe a very rough approximation of the number of columns...
That reminds me, I have to get my VIC-20 back from that guy...
I have to agree. "Clarica" sounds like it has something to do with cell phones...
What possible use can a cookie provide except for exploitation of those that visit your site? For any example you provide I can provide an alternative method for saving info on the web host's drive (secured and with the client's permission of course) that would accomplish the same task.
And then every time people visit your site they have to log in with a userid and password, or a special PIN. What about moving from one page to the next (ie: for a shopping cart)? You could store the state in the URL I suppose. But none of those really buy you any extra security. Infact, they might buy you less. I've seen systems where the userid and password were stored as clear-text in the URL's.
The thing to remember with cookies is that the information comes from the server. In other words, it's information they already have. All cookies give web designers is persistent state. For most sites that persistence only lasts for the current session as well. Cookies are only a problem for the ultra-paranoid.
This reminds me of my favorite RMS quote:
Those who benefit from the current system where programs are property offer two arguments in support of their claims to own programs: the emotional argument and the economic argument.
The emotional argument goes like this: ``I put my sweat, my heart, my soul into this program. It comes from me, it's mine!''
This argument does not require serious refutation. The feeling of attachment is one that programmers can cultivate when it suits them; it is not inevitable. Consider, for example, how willingly the same programmers usually sign over all rights to a large corporation for a salary; the emotional attachment mysteriously vanishes. By contrast, consider the great artists and artisans of medieval times, who didn't even sign their names to their work. To them, the name of the artist was not important. What mattered was that the work was done--and the purpose it would serve. This view prevailed for hundreds of years.
According to RMS, those Apple employees should just be happy that the work was done, and shouldn't be concerned with having their name on it. After all, getting credit for something is like saying you own it, and we all know* that owning software is evil.
---
* yes, that's sarcasm
Incidently, I do agree with you that the definition of "hacking" probably shouldn't include any moral elements. Hacking does imply something that is clever, however, and creating unmaintainable code is most decidedly un-clever.
There already seems to be a gray area here... Persoanlly, I think that if your code is completely non-maintainable and illegible, then it's automatically disqualified from being "brilliant", and hence is sounds more like a kludge than a hack to me. Maintenance is 98% of the typical long-lived software project.
Because chording is slower. While I'm hitting backspace, my left hand can be getting ready to hit the next key. If I use ^H I either have to use both hands, hence slowing me down, or I have to use one hand to chord, which is painful and awkward (especially for a key like H which is almost as far away from both control keys as possible).
BTW: If you don't believe that my left hand can be getting ready to hit the next key, try typing "stargaze" and "shamrock". Most people can type the latter significantly faster, because it requires alternate hands when typed on a qwerty keyboard.
The "zenning out" you refer to is usually called "flow" I believe. It's discussed in Alan Cooper's "About Face" on page 127. Cooper is rather adamant about creating UI's that don't disrupt the flow. He's very much against useless modal dialogs, for example.
I'm a VI user myself (actually, I usually use VIM). I don't have any problem with where escape and control are on PC keyboards. The escape key is farther, but not much farther. It's on the corner though, which I find makes it much easier to find by feel. When I go to press escape, I just reach for the corner key with all four (non-thumb) fingers, without fear of hitting anything else. The key isn't bigger, but by using several fingers it balances out.
As for control, it's also in the corner on PC keyboards, which I find makes it relatively easy to find by feel. I do agree that caps-lock in in an exceedingly dumb place though. Having it be a littel "chiklet" up by the caps-lock light would probably be more appropriate.
Regarding your complaint that there should be control, alt and "vanity" keys on the right side of the keyboard too... huh? Every PC keyboard I've seen *does* have the shifting keys on both sides of the keyboard. That's part of the reason the space bar is smaller than it used to be. You can't have it both ways.
Finally you had several complaints about the cursor-key layout. I think the upside-down T is pretty good, actually. Moving my middle finger between the up and down keys is very easy, and I can keep fingers on the left and right keys in either position. I rarely want to go both up and down simultaneously...
Again, by being segregated, the cursor keys are fairly easy to find by feel, and the distinctive shape makes it easy to position your fingers on the keys without looking away from the screen. It's much better than the cluster of six keys above them.
Sun keyboards also have annoyingly small backspace keys. Maybe you never make a typo, but I use backspace quite frequently...
Finally, as for mice, yes there is a big lag involved in switching between the keyboard and the mouse. Well designed software takes that into account, and has keybindings wherever possible. Unfortunately, many X apps are worse about this that their MS-Windows counterparts.
Ideally, software should be designed such that if you *have* to use the mouse (ie: in a paint or drawing program), you should be able to keep your hand there for a long time, and have one-key accelerators for your non-mouse hand to use. Deluxe Paint III on the Amiga did this. Actually, Star Craft does this as well...
My big keyboard nit (this is actually a software one): why the #&!! are the default behaviour of the backspace and delete keys on Linux (at least, on all the distributions I've used) so screwed up? Backspace should be ^H, and delete should be ^?. Yes, I know how to fix it, but the default is dumb, IMHO.
Java applets already have a restriction preventing them from opening connections to any machine except the one they came from. To get around this applets must be signed, and you would have to accept the corresponding certificate.
As others have pointed out, adding such restrictions to images and frames would break too many things. If you really want to, you could try adding such a feature to Mozilla, and see what it does.
In any case, I believe this Comet Cursors thing uses a plugin. If that's the case, it could probably just go and open a socket connection on its own. The same goes for any other plugin or Active X control (which is what I assume you meant by "embedded controls").
Postscript isn't going away any time soon, and the glitches you mentioned will be eradicated over time in classic open-source style. The turing-complete problem isn't a horrible problem because the postscript-generating program can limit itself to generating postscript commands that are known to produce predictable results.
I'm not suggesting that PostScript go away. I don't mind printers understanding it. I just don't think it should be the API that application programs use to talk to the printer.
And no, nothing, not even the open source movement, will erradicate the halting problem. It's been proven mathematically.
Finally, saying that an application program can "limit itself to generating postscript commands that are known to produce predictable results" is similar to saying OS's shouldn't have memory protection because applications can "limit themselves to accessing memory in a ways that are known to produce predictable results". The fact is, many applications don't. Many times I've tried to print things "2up", only to find that the application was generating postscript that circumvented it. A proper printing API wouldn't allow that.
And even if it weren't for these problems, does it really make sense that application writers should need to learn PostScript in order to print? That's like requiring people to learn the X protocol if they want a GUI. There should be an actual API something like Xlib, or at the very least a simpler page description language, like one based on XML.
There is no technicial problem in the following code, style is another issue, but compile it with 'gcc -pedantic -Wall' and gcc will complain "`x' might be used uninitialized in this function."
Yes, but if you look at the code in ghostscript, you'll see that it isn't a trivial situation like that. Most of the places I looked at had switch statements inside loops. I haven't investigated it in great detail, but it looked suspiciously like it was scanning though the input, and initializing certain variables in only some of the cases in the switch statement. If that's correct, then it's possible that invalid input would make ghostscript use the values of uninitialized variables.
Even if the code was correct at one point, I doubt that it is now. Bad style generally leads to errors, especially in programs that are old and have many maintainers.
Don't get me wrong, I'll take Linux over Windows any day, but I'd have to say that one of Linux's weaknesses is no real support for system-wide printer drivers. Most Linux and Unix apps assume that you've got a PostScript printer. If you don't you have to use something like ghostscript to act as a filter for your printer.
There are numerous problems with this setup. For one thing, ghostscript is complicated, rather badly written in places, and crashes far more often than I'd like. To get an idea of how hokey the ghostscript code is, try compiling it with '-pedantic -Wall'. You'll see tonnes of "variable x may have not been initialized" warnings...
Using PostScript as the API for communicating printers is just a bad idea. PostScript is Turing complete. That means there's all sorts of analysis you simply can't do to the stream sent to the printer. It isn't even guaranteed that the printer will ever finish (the halting problem). It would be a lot simpler if there was a drawing API, sort of like a subset of X, but for paper. Or heck, an XML-based page description language. Just make sure it isn't Turing complete.
If Linux actually had decent printing system, then WordPerfect wouldn't need its own. As it is, printing is one of Linux's biggest weaknesses, IMHO.
Having lived in both San Francisco and Toronto, I'd have to say your comment is incorrect. Canadian culture is quite different from the US. Those differences just aren't as easy to see without actually being immersed in both cultures, because the media in both places are very similar. And I'd have to say Toronto is more diverse than SF. Canadian culture as a whole is certainly more tollerant of diversity than American culture.
I once read about a study that was done in towns that lie on the border between the US and Canada. The study was essentially an opinion survey. They then checked to see if there was any significant difference between the Americans and the Canadians. A significant difference was found.
Perl is probably the most widely used language for CGI because of the lemming effect. If it had to do with effeciency of the resulting code, perhaps a compiled language like C or C++ would have been used. (though as others have pointed out, interpreted languages like Perl and Python will be just as fast as C or C++ for things that spend most of their time doing string processing) If rapid development was the goal, then there are other languages (like Python) that are arguably easier to write in than Perl. Perl is the most popular now because it was the most popular before.
Perl was the first language to become popular for CGI. Python didn't exist at the time CGI first appeared (AFAIK), and the web was young. Enormous commercial websites didn't exist, so efficiency wasn't a huge issue. CGI writers were mostly just trying to do "cool stuff", not develop large maintainable systems, so ugly but relatively easy to write Perl code was perfect.
Because virtually all of the early CGI scripts were written in Perl, Perl became the de facto CGI scripting language. Now most CGI programs are written in Perl simply because that's what everyone else does. If you go to the bookstore and look for books on CGI, most will be the Perl variety. Do a search on Google for "cgi script" and the first link has "Free Perl CGI Scripts" in the description. There are even some servers where they'll only let you have CGI scripts if they're written in Perl. People think "Perl" when they hear CGI.
This is a vicious cycle of course. The more popular Perl is for CGI, the more first-time CGI writers assume you need to use Perl, so the more popular it becomes. Perl is to CGI as MS Windows is to computers. Newbies think that's what they need to use, because of sheer numbers. The typical CGI newbie doesn't know you can use anything other than Perl, in the same way that the typical PC newbie doesn't know that their PC can run an OS other than Windows.
For the record, my first CGI program was written in Perl. I was going to write it in Rexx, but the server I was going to use only allowed Perl. All of the CGI programs I've written since have been written in Python (and they typically run on machines where I am root).
An idea I heard about a while back was to prevent MS from buying anything to produce their software except labor, and hardware. They wouldn't be allowed to buy other companies (or divisions of other companies), nor could they make "deals". They also wouldn't be allowed to "distribute" software written by others. Essentially they would be forced to innovate, rather than buying the innovations of others.
MS is pretty sneaky, so care would have to be taken to make sure they don't find a loophole. The best way to do this is probably to say MS may not spend any money except on the things in a carefuly chosen list (which would include employee salaries, hardware, etc.).
MS could then no longer buy out all of the competition, nor could they act as simple distributors for other companies products. They'd actually have to create things on their own. It would be hard for MS to make this look like "punishment" from a PR point of view. After all, they say they want "the freedom to innovate", and this wouldn't prevent them from innovating. If MS really was as innovative as they claim, such a sentence wouldn't hurt them significantly. I'm willing to bet it'll damage them quite seriously however.
Build a robot with your old PDA as its brain. Most PDA's have a reasonably powerful CPU, with a very low power consumption - perfect for a little robot. You might even be able to do this without opening the PDA if it's got a connector for a cradle.
Huh? Are there really people who believe that "creativity thrives on unlimited freedom"? Sorry, but it ain't so. I've always felt that the truly creative are those that are able to work within the rules, and still come up with something original and interesting.
There are many art forms that impose strict rules and structure including many forms of poetry (sonnets, haiku, etc.), most forms of music, dance and architecture. Without some sort of structure these become babbling, noise, convulsions and heaps of rubble. Structure is good. The truly creative can work within structure to create beuty. The non-creative wannabe's whine about being repressed.
I think the meta-moderation is going too far. Multiple moderators should be enough to balance things out. Giving people more moderation points would server to balance more. If I thought a moderation was unfair, and I had moderation points, I'd "counter-moderate" the post.
Dummy moderation controls is probably a bad idea. Only the obsessive will use them. The people who actually care about doing something useful simply won't use the controls if they think they'll probably just be ignored.
Also, I have to say that the "auto moderation" thing is very nasty. Just because someone posted a good comment or two, it doesn't mean all of their comments are good. Let each comment stand on its own. Likewise, there is the ongoing debate about whether AC posts deserve to start at 0 or 1.
How about this:
1. Increase moderation points, and don't make it quite so hard to become a moderator.
2. Remove the auto-moderation thing. All comments start at 1 (even AC posts).
3. Add a "killfile"-like feature where each user can artificially bump-up/down the scores of posts on a per-poster basis. Each person's "killfile" will only affect the scores of posts as they read them. So if you think AC posts deserve to get 1 less than everyone else, you'd put "Anonymous Coward -> -1" in the "killfile". You could even put this in everyone's killfile by default so the behaviour would be identical to what it is now (by default). If this is done, I think the score range should be chnged to -5 to 5 (instead of the current -1 to 5) and the "killfile adjustment" range should be -10 to 10 (so I can bump a -5 to a 5 or vice-versa if I think the poster is deserving of it).
3. From the main page, if the user has moderation privledges, add a "Moderate" link in addition to the "Read More" link. The "Read More" link will be for reading only, no moderation controls. The "Moderate" link will bring one to a special moderation page.
4. The moderation page would have moderation controls of course. It would also not show the scores, and sorting by score wouldn't be an option while moderating. And your killfile wouldn't have any effect in moderation mode either.
5. Finally, I don't think it should show who posted the comment. I really think it should be a meritocracy, and showing people's ID's to the moderators prevents that. I realize that everything having no ID could get confusing (especially in conversations), so I think for each person that posts a comment in a discussion area, there should be a new "fake ID" generated. This fake ID would be used for all of their posts on the moderation page. Fake ID's would not be persistent. In one discussion area my fake ID might be "Hoopy Cow" in another is might be "Green Frood". That way you could still follow conversations in a discussion area, but the identities of the posters would be disconnected from what they are in "Read Mode" and from all other discussion areas. Note also that Anonymous Coward would be treated as another user, so it wouldn't be obvious that AC posts were from AC's. Again, comments would be judged more on content rather than who posted them. Oh, and almost forgot. Signatures wouldn't be shown in moderation mode of course.
I realize that someone could go and open both "read" and "moderate" to find out who made various posts. It would be possible to disable moderation of discussion areas the user has alredy read, but then they could just read as AC, and moderate as themselves. I think that the above suggestions would at least reduce the chance of people moderating based on identity rather than content though. Having the moderation controls shown right there with the poster's identity just makes it too easy to say "hey, I know that guy" and +1, or "oh, I hate him" and -1.
Oh, and some gratuitous other suggestions:
- Make plain-old-text really plain text.
- spell checking!
- link checking!
- better conversion of busted URL's ("relative" URL's should probably be patched up to simply add a leading "http://", as that seems to be the most common error)
- what's with those "parent" links? I can never figure out which one is really the parent when I'm in flat mode.