the compound nouns in Scandinavian/Teutonic languages are examples of bad linguistic design
No. Software is more flexible than people's habits, and as there's already a standard way of dealing with the problem (the soft hyphen, which is a defined part of the UNICODE spec) there is no reason whatsoever for software to not handle it. Anything else is just stupid linguistic imperialism. (Plus, sometimes it's really useful to be able to say where to hyphenate long wordlike-entities, even in English.)
[S]ome versions of sendmail automatically go to DNS and ignore the resolver libs. Why, I don't know; that's what gethostbyname() is for, after all...
It's probably because ignoring gethostbyname() lets you avoid having to use masses of processes to do asynchronous lookups. If you have a reentrant version of gethostbyname, you can use threads instead of processes, but it's still far more expensive than a real asynch lookup that lets the main select() call tell you when the message comes back. I'd even go so far as to advise everyone to use this sort of technique in your own programs, except that discovering exactly how names are resolved is far gnarlier than it appears to be at first glance; IIRC, there's one API configured on at least one platform which is actually done through just stating which dynamic library knows how to handle the whole thing (including which other configuration files to read). Having worked in a place where NIS+ was the way names were resolved, and corresponded with a place that used LDAP (!) bypassing creaky old gethostbyname() (or it's not-quite-improved-enough younger brother, getaddrinfo(), which is at least safely usable from multiple threads at once) is a bad idea. Alas.
Re:Backgrounds of the PHP developers.
on
PHP 5.1.0 Released
·
· Score: 1
Personally I think the use of the "global" keyword is something PHP got right and every other language (except ruby) got wrong. [...] PHP got it right, perl got it wrong.
Tcl is another language that uses that approach (i.e. variables are always local unless explicitly declared otherwise). You're absolutely right that Perl got it wrong, even though they later hacked in something that goes some way towards correcting the mess (which is itself absolutely typical of Perl...)
Bah! Turing-completeness just buys you a seat in the game. It's practical power that really matters, and that's something else entirely. After all, unlambda's all very well but it's nice to be able to open sockets, build and run GUIs, render webpages, query databases, etc. It's the libraries (and the cool things you do with them) that really count.
So, you think strong typing and heavy compilation (with all the attendant futzing around with silly little converter classes and twiddling your thumbs while the whole lot grinds through the compiler again) are necessary for OO-in-business-scenarios? Interesting. Can't say I agree though. Highly typed languages are overrated (and PHP is definitely not best-of-breed in the light-typing end) and compilation to native code is not so useful as you might think; there's usually a better level of abstraction to use as a compilation target (i.e. which allows for very fast compilation while providing decent speed), assuming you're not writing some real CPU-heavy code like image manipulators. But then most web-apps aren't about that; they're "parse the input, diddle with the DB, possibly contact some other service, format the output". All the scripting languages are good choices for that sort of thing, though some are better than others for mundane reasons. Reasons like "the natural way to write code is the secure way to write code!" But you take your pick according to your personal preferences (or sometimes those of your boss, alas.)
I must say though, that a plugin shouldn'be able to crash Firefox itself, although it does.
Don't hold your breath waiting for that to happen. Plugins are inherently a potential source of trouble since they're "plugging in" extra code into the browser (which is how they support their functionality, of course) and if they've got a bug the crash can take out the browser itself. While it is possible to write plugins such that virtually all the plugin code actually runs in another process (some plugins work this way) they cannot run entirely in a separate process, and so cannot be totally isolated.
FWIW, this isn't a Firefox issue. It's just a fundamental problem with all plugin-based architectures (Windows is particularly infested with this sort of trouble, given that it's all founded on COM, which is itself the same sort of thing as a plugin arch...)
postgresql->oracle is the best combo in the book. Every project my company has touched in the last 7 years started with postgresql and if we ran into one of the limitations in postgresql, switching to oracle was painless.
I'd agree with that, and point out that SQLite is there for people who come in at the small end of PostgreSQL. I'm told (by the SQLite author no less) that upgrading to postgresql is easy when you want to take that step. And it is public domain.
Between them, sqlite->postgresql->oracle offer a full database solution for everything from "I want a better config file for my personal scripts" to "I have to run a mission-critical database for a Fortune 500 corporation", and you can't say much fairer than that.
Even megastars from 2 years ago are failing to draw a crowd (Gavin Rossdale of Bush drew only 40 people in Milwaukee two weeks ago).
Who? Nobody can be a megastar if I've not heard of them, OK? I know I don't put a lot of effort into tracking what's going on in the pop scene, but if someone's a real megastar that won't matter; I'll still have heard of them.
I agree, there are no military applications. Irrespective of what you think about current military activities.
You don't want the soldiers to become fearless, because if they do, they might say: "This war is wrong. I used to be too afraid to do anything about it, but now I suddenly feel fearless, and will get the heck away from here !"
To be fair, fear is also a part of what keeps soldiers from taking stupid risks in a firefight. Whatever else you say about the army, they don't want lots of casualties; saps morale of the troops themselves, and of the folks back home too. And to cap it all, it costs lots of money to replace the injured/killed person (a quick search indicates that training costs can run to $250k a head for someone who knows what they're doing, and that's only one of the costs involved with soldiering). Oh, and who knows how much the gene therapy to make them fearless would cost?
In any case, whether or not you think wars are justified, you don't want fearless soldiers. They'd be horribly expensive.
I know it's popular to hate lawyers, so what I'm about to say will probably burn my karma to cinders. But, the simple truth is, it's not the fault of lawyers. They're working within the system, getting paid by clients to do what they do. You want less patent lawsuits? Reform the patent system. Don't burn lawyers at the stake.
Oh, I'd keep the idea about burning the lawyers around. But remember why the system is broken; make sure you use the politicians as the fuel for the lawyer-burning.
You are assuming that an Unregulated Free Market will necessarily remain free - note that you're arguing against regulation, since that requires the existence of an organization capable of enforcing the regulations - which is an interesting piece of dogma, but cynic that I am, I don't believe it. And I think there's more evidence (look at the history of the Sherman Antitrust Act) to back my position than yours.
I'm just figuring that most IRC fans haven't figured out to look in comp.embedded, or comp.linux. Or to sign up to some mailing lists.
Just like I'm figuring you don't know the difference between talking to someone face-to-face and writing letters to them? When you're operating at the real cutting edge, the immediacy of a chat system like IRC (or Jabber) can't be beat; the alternatives to it are really either telephone conferencing or real physical meetings, both of which can get really expensive to set up and run on a regular long-term basis.
You are aware that that's exactly what Eolas are doing? (I've met the people from Eolas, and they're smart regular technical people and not patent leeches, err, lawyers.)
The academic grids aren't free. It's just that it's the Principal Investigators who get the bill, and not their jumpsuited minions, err, research students.
There I was thinking (from the headline) it was about unified source code pretty-printer and highlighter. Instead, we get boring stuff about changing the exact look of all the icons all over again. Hey guys, this is just about keeping lame-ass graphic designers off the streets, and not about making OSS better for developers...
Reason 1: because it's an inter-page ad. You know, the ones that use javascript to make it so that you don't go to the page you were expecting. If you're going to put an ad in there, you'd better be honest about the whole thing.
Reason 2: because the ad server is so slow that loading the rest of the page is massively delayed even over your nice gigabit connection to the network backbone. Cooling my heels waiting for some silly random server to serve up content I don't want just so I can see the content I do want (and which is already downloaded, to add insult to injury) isn't my style.
My preferred method of blocking? Editing/etc/hosts to make my system think that the offending systems have IP addresses in the 127.*.*.* netblock.
(I also somehow managed to get my browser to not have any Flash support. That suits me just nicely!)
Having travelled through far too many airports this year, the problems are two-fold. Firstly, bad design. Too many airports, especially but not uniquely in the US, have their security purely as an afterthought and that leads to horrible backlogs. Secondly, at check-in there's a problem with a higher proportion of people with inherently complicated problems with luggage or ticketing.
If you can, avoid O'Hare for international departures. Probably best to avoid Detroit and Logan (especially Logan!) too. Within Europe, Heathrow and De Gaulle both have their own special suckiness, and Zaventem's no place to get a cab. (There may be worse airports, but not that I've visited recently.)
Slashdot Mirror
Bah! Turing-completeness just buys you a seat in the game. It's practical power that really matters, and that's something else entirely. After all, unlambda's all very well but it's nice to be able to open sockets, build and run GUIs, render webpages, query databases, etc. It's the libraries (and the cool things you do with them) that really count.
So, you think strong typing and heavy compilation (with all the attendant futzing around with silly little converter classes and twiddling your thumbs while the whole lot grinds through the compiler again) are necessary for OO-in-business-scenarios? Interesting. Can't say I agree though. Highly typed languages are overrated (and PHP is definitely not best-of-breed in the light-typing end) and compilation to native code is not so useful as you might think; there's usually a better level of abstraction to use as a compilation target (i.e. which allows for very fast compilation while providing decent speed), assuming you're not writing some real CPU-heavy code like image manipulators. But then most web-apps aren't about that; they're "parse the input, diddle with the DB, possibly contact some other service, format the output". All the scripting languages are good choices for that sort of thing, though some are better than others for mundane reasons. Reasons like "the natural way to write code is the secure way to write code!" But you take your pick according to your personal preferences (or sometimes those of your boss, alas.)
FWIW, this isn't a Firefox issue. It's just a fundamental problem with all plugin-based architectures (Windows is particularly infested with this sort of trouble, given that it's all founded on COM, which is itself the same sort of thing as a plugin arch...)
Between them, sqlite->postgresql->oracle offer a full database solution for everything from "I want a better config file for my personal scripts" to "I have to run a mission-critical database for a Fortune 500 corporation", and you can't say much fairer than that.
In any case, whether or not you think wars are justified, you don't want fearless soldiers. They'd be horribly expensive.
Can I plant B-Trees to get that saving? Or does it have to be R-B or AVL trees? Would a DAG do instead?
entity vs. nonentity
That's the IBM vs. SCO case, isn't it?
You are assuming that an Unregulated Free Market will necessarily remain free - note that you're arguing against regulation, since that requires the existence of an organization capable of enforcing the regulations - which is an interesting piece of dogma, but cynic that I am, I don't believe it. And I think there's more evidence (look at the history of the Sherman Antitrust Act) to back my position than yours.
GORMless
You are aware that that's exactly what Eolas are doing? (I've met the people from Eolas, and they're smart regular technical people and not patent leeches, err, lawyers.)
The academic grids aren't free. It's just that it's the Principal Investigators who get the bill, and not their jumpsuited minions, err, research students.
"Great minds think alike; hurried developers make similar mistakes."
There I was thinking (from the headline) it was about unified source code pretty-printer and highlighter. Instead, we get boring stuff about changing the exact look of all the icons all over again. Hey guys, this is just about keeping lame-ass graphic designers off the streets, and not about making OSS better for developers...
Reason 1: because it's an inter-page ad. You know, the ones that use javascript to make it so that you don't go to the page you were expecting. If you're going to put an ad in there, you'd better be honest about the whole thing.
/etc/hosts to make my system think that the offending systems have IP addresses in the 127.*.*.* netblock.
Reason 2: because the ad server is so slow that loading the rest of the page is massively delayed even over your nice gigabit connection to the network backbone. Cooling my heels waiting for some silly random server to serve up content I don't want just so I can see the content I do want (and which is already downloaded, to add insult to injury) isn't my style.
My preferred method of blocking? Editing
(I also somehow managed to get my browser to not have any Flash support. That suits me just nicely!)
Having travelled through far too many airports this year, the problems are two-fold. Firstly, bad design. Too many airports, especially but not uniquely in the US, have their security purely as an afterthought and that leads to horrible backlogs. Secondly, at check-in there's a problem with a higher proportion of people with inherently complicated problems with luggage or ticketing.
If you can, avoid O'Hare for international departures. Probably best to avoid Detroit and Logan (especially Logan!) too. Within Europe, Heathrow and De Gaulle both have their own special suckiness, and Zaventem's no place to get a cab. (There may be worse airports, but not that I've visited recently.)