Other people have mentioned that Regular Expressions aren't native in C/C++, which is correct, but there are regular expression packages. You may have to say re_compile(re); re_execute(re,str); or whatever the syntax is, but it can be done relatively easily. The fact that it is not built in does not make it slower.
You do not know what your C functions are for regexps? I, and every other perl hacker out there can tell you about =~ and// and other perl-isms for matching and manipulating regexps in relatively greater detail.
Thanks for agreeing with my point - it's time to implement that I was talking about, not performance increase as a result of using C, and compared to having to get out the manpages to look up your regexp functions *let alone* the depths of ODBC's SQLGetInfo() call, you're imposing a far greater workload on the poor coder. For a probably-not-vast increase in performance.
The speed is based on the quality of implementation.
Impossible generalisation, but remarkably obvious.
I suggest you try out a few non-C things for a change.
There is a perl compiler but depending on how you compiled your perl installation you get either a static (*LARGE!*) binary or a dynamic one (shared library loading overhead).
As a general rule you really don't want to be fork()ing a new process (especially involving loading a binary off disk) for *every* hit to a dynamic page. That's partly why some of the NT benchmarks shine, because people are using ASPs (basically, DLLs) using ISAPI instead of real CGI. This is analogous to mod_perl, where the perl interpreter stays up for more than one hit (IIRC).
Of course, the main reasons for using perl is that string manipulation is/way/ easier - you're flinging regexps around with mere punctuation all in one line; not even tcl gets close, having to call functions to do regexp matching. And as for C or C++... eek. Excessive curly-brace overload!
Also, apache is modular, perl is modular. You want php3 scripting, easy done. You want perl, also easily done. You want perl modules to access RDBMS behind the scenes, very easily done. (Likewise php, for that matter.) Trying to incorporate that sort of thing into C or C++ might result in a speed increase of execution, but if you're still loading entire binaries off disk each time, it's not likely to be that significant and you've got to crank out the whole API for your backend RDBMS (ie CT-lib for Sybase Open Client, ODBC for ODBC access, OCI for Oracle, whatever) which is a large development investment overhead for stuff-all performance increase.
Yeah. The problem with slashdot is that moderators don't know to leave an article alone - if it in any way duplicates another point you get marked down as "redundant" to the extent that you feel you can't even agree with things. It's a permanent state of "must say something more insightful / funny". If there was more leaving of things alone, neither your problem nor my opposite would happen...
(Also, I wonder exactly how folks approach increasing an individual comment's score - it appears to be a combination of dB of laugh caused or count(#links) in it...!)
Still, what is a good memory worth if you can't think properly?
It works both ways: what use is being able to think, if you have nothing to think about? I suspect there's a kind of continual process where thought becomes knowlege anyway - I know that's roughly how I work, in that once I've thought my way through a few steps to a conclusion, I remember that jump as a more fundamentally "viable" thing. YMMV of course;)
I've heard people say that Cancer and AIDS/HIV are nature's way of dealing with overpopulation.
Such people are merely ignorant twerps trying to press their moral agenda on you. I think you can safely ignore such simplistic approaches. Ask one at random to prove that it's because of overpopulation that AIDS/HIV got started, and whether they wouldn't mind dropping down dead in order to obviate the overpopulation "problem" themselves. That might at least cause an interesting reaction;)
Don't even think about y2k problems, get a beach house and a towel:)
The way to stop kiddies looking at pr0n is not to stick some poxy Net Nanny in the way that obliterates just about everything (Scunthorpe!) but to educate them not to want to go looking for it.
Similarly if teachers have to resort to banning calculators and Palm Pilots (eek!) because of the potential for fun with IrDA, they're not doing their jobs properly. They should be inspiring the kids to learn what they're teaching, not play network-reversi at each other or whatever.
As for exams, well there shouldn't be line-of-sight, or maybe they should be blocked off with a password for that sort of thing.
*Sigh* The problem is, anarchy is/fine/ by me -and I tend to support it way over any government involvement- if it all works out generally nicely - but as soon as you get real blood on the carpet, it's gone way too far.
I see traces of Snow Crash here. The interaction between "cyberspace" and the real world, the role of anarchy... strange!
There's something else that must be considered here: how exactly did they test the modem in order to establish the speeds of pings?
If you want a realistic idea of what to expect, you should have a couple of phone lines installed and an "ordinary" modem and the new thing both in / attached-to the same machine, and try pinging simultaneously, at the same speed connection, for a long while. Then you flip the modems over to the other line, and try again.
What's the point in something that manages to respond faster to ICMP pings anyway - is it an established factoid that games like quake2, say, do a lot of pinging compared to real data?
Say what? Where does the Linux community come into it? It's not mentioned on the nerdperfect competition page, nor is slashdot only the linux community. NT *does* take too long to boot. So do a lot of things, including my Debian installation because I've got loads of junk in the startup scripts that I've not purged, more twerp me.
It would be more to the point if they had a 'world's slowest bootup of an active system whatever the OS' competition, IMO...
Just an idea too... will it run on my palm pilot?;)
Semi-serious point, anyway. It occurred to me that there are one or two applications out there where (a) the distinction between RAM and disk is blurry, to say the least (b) things stay in a running or runnable state between power-cycles.
I've acquired Zipdisks; Eros is going home tonight to be played with:)
> Opponents of NCs always point out that before PCs existed everything was done on NCs connected to micros or mainframes, and thus a move towards NCs would be a step backward which is doomed to fail.
Yeah. But then again, it was pointed out many years ago while I was at uni that the industry swings back and forth between integrating FPUs on CPUs, between putting networking, sound, graphics (in particular) into the CPU... so just because it's the same as something that's already happened doesn't mean that it's wrong, or "doomed to fail".
I'm not sure where you're getting at with the >$200 restriction though. Unless it's that a lot of folks have already gone out and got far more expensive toys than that to do as little as that with. But something like $200 instead of $1400 is definitely lowering the boundary between the information-haves and info-have-nots, which is probably a good thing.
> Domain names are the addresses computer users enter to access a particular Web site. They come between "www" and "com," "net" or "org."
Aaaaaargh. They come before ".us", don't they?:]
> the Clinton administration opposes the legislation, arguing that the best place to resolve domain-name disputes is in the courtroom.
Like how are they arguing? I see no links, no discussion... I don't think taking everything to court is a particularly productive way of settling every little dispute under the sun, although it might be considerably better than domain-names being bought-out for ever-escalating sums. Can't they just agree to let some friendly InterNIC-derivative have absolute control over what domains an organization is allowed?
Looks like pretty lousy "reporting" (CNN? I thought they were going up in the world?) and some bad ideas..
>I just recently took a class in X-Windows. ...etc. Yes, well Motif isn't exactly the best way of doing everything, or so I've heard (comments about square roots and roman numerals come to mind:)
> As to your pity being directed at us poor VB programmers, thanks but no thanks. VB has an amazingly productive UI. It sure beats the hell out of 4 xterms and writing my own make files. That's up to the individual, I think. But give glade a blast before writing the non-VB world off as "4 xterms and writing your own Makefiles"!;]
*Cough*. From experience, VB3 produced executables with only dependencies being vbrun*.dll.
Let's just say that when I did pre-University work experience, I learnt & hacked VB3 and above. When I got to Uni, I was doing C, C++, ML,... and now I'm working in the IT industry (surprisingly enough) I'm not doing that much actual coding. Perl scripting is prevalent, java does the rounds, and the mainstay is C.
Actually there are two flip-sides to what you say. Some thoughts that come to mind: It's all very well everything being "consistent" but if you're consistently Microsoft, you're consistently closed-source. There comes a point where the warm fuzzies of "building" on existing APIs freeze their bits off and you realise you're just clicking-in more bloatware with tie-ins to Windoze, to Micro$loth, to your particular development environment ("what vbrun300.dll?"). Does VB do containers for its controls like java, or like gtk?
Is COM really common? Is it free? Where are the implementations on Unix boxes? Do people even want such implementations?
You have a *major* misconception of how things operate in your phrase "Linux can't agree on the guts behind their GUIs", as though "Linux" were a company deserving of "their". To cut to the quick, don't do it:)
There is always glade, an open-source development environment for gtk+ pushing out C, C++ or Ada code (Perl to follow, at the last count). That's quite close to a "visual" environment, if you think such things are cool. That and I suspect you've never done any coding in xemacs, either. Also you've got gnome/kde the wrong way round, IIRC: KDE is the one that uses some-CORBA,some-proprietary stuff; AFAIK gnome tries to do *everything* through CORBA. (There have been discussions on/. before now about this, I'm sure:)
It's fair enough that 'simple doesn't mean smart people shouldn't use it', but it does make it quite likely that there will be people out here who don't want to use it, quite often because it doesn't do all they want. Developers are and should be prepared to work to get things right.
A micro-story: at my last company, we had one customer from hell, as far as support was concerned. He was writing stuff in VB, using ODBC to connect, and he had a problem. I long-since decided it was a bug in VB, for various reasons (it was:) and decided, amidst the insults, to have a look at his VB project. After looking through it for 15minutes, despite his bad attitude I was thinking "this chap is paid, what, £25k/pa, to spend all day every day looking at this interface hacking this language?!", and I actually felt sorry for the chap!
You're right. I, for one, never need go to http://www.lycos.co.uk/ (where the blasted thing redirects me to - talk about wasting bandwidth with DNS lookups), simply because there's nothing of any interest on there. OTOH, I do prefer to use FTPSearch and MP3search, both of which are not expressly linked off the front page.
To me, it says it all that the actual web-searching bit has been reduced to a tiny box at the top of the page.
On the positive side, I noticed the other day that it seems to have turned into a bit of a portal: the question is, too little, too late?
> Of course, this is not flamebait - I don't necessarily hold the same opinion for animals which have been eradicated by humans.
Out of interest, why not? OK, so we're the only species we know of to have evolved slashdot, but that's no reason to consider humans separately, especially *if* you think we 'evolved' from some other kind of mammal.
> Yes, it was obviously a joke -- however, I'm getting a little tired of the "it's redhat 4.x,5.x,6.x so it must be insecure" attitude paired with the "I've seen that movie hackers so it must be easy to hack" jokes.
I hope you're not suggesting that either attitude was present in my comment...
My point is that (a) a proper ruleset is always a good idea, (b) firewalls are still susceptible to breaches, be it by brute force or some other means (c) information given out makes folks' jobs easier to hack whatever it is, whether they should be able to get to port 23 or 80 or not.
<cynical>Oh look! A Unix workstation! I know how to use those!</cynical>
Then you, sir, are a scumbag and not a gentleman;)
Yes, it's blindingly obvious - it doesn't just have to www.royal.gov.uk, it could equally well be freshmeat.net - there are folks out & abouts who *don't* want it defaced. And if they're the ones providing the machinery to folks, they have the right to protect it. And letting on "it's RedHat 5.2" is probably supplying someone with a little too much. IMO:)
You must be aware you're looking for Debian, with its.deb format and particularly, apt utilities... In particular, this link explains the differences fairly well. Being able to set up a list of sites with unstable and stable distributions and contrb / main / binary / non-free / non-US subdirectories, combine a list of package versions from them all, and then whenever you want to install something, 'apt-get install package' will do the job - INCLUDING dependencies.
At least RPM 3 has the ability to cope with HTTP URLs on the commandline - even if you do have to paste them in by hand....
Slashdot Mirror
Other people have mentioned that Regular Expressions aren't native in C/C++, which is correct, but there are regular expression packages. You may have to say re_compile(re); re_execute(re,str); or whatever the syntax is, but it can be done relatively easily. The fact that it is not built in does not make it slower.
// and other perl-isms for matching and manipulating regexps in relatively greater detail.
You do not know what your C functions are for regexps?
I, and every other perl hacker out there can tell you about =~ and
Thanks for agreeing with my point - it's time to implement that I was talking about, not performance increase as a result of using C, and compared to having to get out the manpages to look up your regexp functions *let alone* the depths of ODBC's SQLGetInfo() call, you're imposing a far greater workload on the poor coder. For a probably-not-vast increase in performance.
The speed is based on the quality of implementation.
Impossible generalisation, but remarkably obvious.
I suggest you try out a few non-C things for a change.
There is a perl compiler but depending on how you compiled your perl installation you get either a static (*LARGE!*) binary or a dynamic one (shared library loading overhead).
/way/ easier - you're flinging regexps around with mere punctuation all in one line; not even tcl gets close, having to call functions to do regexp matching. And as for C or C++... eek. Excessive curly-brace overload!
As a general rule you really don't want to be fork()ing a new process (especially involving loading a binary off disk) for *every* hit to a dynamic page. That's partly why some of the NT benchmarks shine, because people are using ASPs (basically, DLLs) using ISAPI instead of real CGI. This is analogous to mod_perl, where the perl interpreter stays up for more than one hit (IIRC).
Of course, the main reasons for using perl is that string manipulation is
Also, apache is modular, perl is modular. You want php3 scripting, easy done. You want perl, also easily done. You want perl modules to access RDBMS behind the scenes, very easily done. (Likewise php, for that matter.)
Trying to incorporate that sort of thing into C or C++ might result in a speed increase of execution, but if you're still loading entire binaries off disk each time, it's not likely to be that significant and you've got to crank out the whole API for your backend RDBMS (ie CT-lib for Sybase Open Client, ODBC for ODBC access, OCI for Oracle, whatever) which is a large development investment overhead for stuff-all performance increase.
Yeah.
The problem with slashdot is that moderators don't know to leave an article alone - if it in any way duplicates another point you get marked down as "redundant" to the extent that you feel you can't even agree with things. It's a permanent state of "must say something more insightful / funny".
If there was more leaving of things alone, neither your problem nor my opposite would happen...
(Also, I wonder exactly how folks approach increasing an individual comment's score - it appears to be a combination of dB of laugh caused or count(#links) in it...!)
Still, what is a good memory worth if you can't think properly?
;)
It works both ways: what use is being able to think, if you have nothing to think about?
I suspect there's a kind of continual process where thought becomes knowlege anyway - I know that's roughly how I work, in that once I've thought my way through a few steps to a conclusion, I remember that jump as a more fundamentally "viable" thing. YMMV of course
I've heard people say that Cancer and AIDS/HIV are nature's way of dealing with overpopulation.
;)
:)
Such people are merely ignorant twerps trying to press their moral agenda on you. I think you can safely ignore such simplistic approaches.
Ask one at random to prove that it's because of overpopulation that AIDS/HIV got started, and whether they wouldn't mind dropping down dead in order to obviate the overpopulation "problem" themselves. That might at least cause an interesting reaction
Don't even think about y2k problems, get a beach house and a towel
I'm not sure that you can - didn't your contract stipulate that things invented were your company's intellectual property?
Of course, if your old company is one to get sued stupidly then it could be interesting...
Me, I think this is a naff idea, unfortunately.
The way to stop kiddies looking at pr0n is not to stick some poxy Net Nanny in the way that obliterates just about everything (Scunthorpe!) but to educate them not to want to go looking for it.
Similarly if teachers have to resort to banning calculators and Palm Pilots (eek!) because of the potential for fun with IrDA, they're not doing their jobs properly. They should be inspiring the kids to learn what they're teaching, not play network-reversi at each other or whatever.
As for exams, well there shouldn't be line-of-sight, or maybe they should be blocked off with a password for that sort of thing.
If that's the case, the thing to ascertain is what the default value of the check-box is.
;)
Me, I don't use realplayer, I use grip for playing CDs and theatres for movies...
I think the problem with that is that Unisys need to prove that you *had* the GIF, don't they?
:)
rm would appear to be a fantastic way to sort out one's conscience on this sort of thing...
> and child porn spammers deserve an espcially painful death.
Really? For what is death *deserved*?
Methinks thou trollest.
*Sigh* /fine/ by me -and I tend to support it way over any government involvement- if it all works out generally nicely - but as soon as you get real blood on the carpet, it's gone way too far.
The problem is, anarchy is
I see traces of Snow Crash here. The interaction between "cyberspace" and the real world, the role of anarchy... strange!
There's something else that must be considered here: how exactly did they test the modem in order to establish the speeds of pings?
If you want a realistic idea of what to expect, you should have a couple of phone lines installed and an "ordinary" modem and the new thing both in / attached-to the same machine, and try pinging simultaneously, at the same speed connection, for a long while. Then you flip the modems over to the other line, and try again.
What's the point in something that manages to respond faster to ICMP pings anyway - is it an established factoid that games like quake2, say, do a lot of pinging compared to real data?
Say what?
Where does the Linux community come into it?
It's not mentioned on the nerdperfect competition page, nor is slashdot only the linux community.
NT *does* take too long to boot. So do a lot of things, including my Debian installation because I've got loads of junk in the startup scripts that I've not purged, more twerp me.
It would be more to the point if they had a 'world's slowest bootup of an active system whatever the OS' competition, IMO...
Just an idea too... will it run on my palm pilot? ;)
:)
Semi-serious point, anyway. It occurred to me that there are one or two applications out there where
(a) the distinction between RAM and disk is blurry, to say the least
(b) things stay in a running or runnable state between power-cycles.
I've acquired Zipdisks; Eros is going home tonight to be played with
> Opponents of NCs always point out that before PCs existed everything was done on NCs connected to micros or mainframes, and thus a move towards NCs would be a step backward which is doomed to fail.
Yeah. But then again, it was pointed out many years ago while I was at uni that the industry swings back and forth between integrating FPUs on CPUs, between putting networking, sound, graphics (in particular) into the CPU... so just because it's the same as something that's already happened doesn't mean that it's wrong, or "doomed to fail".
I'm not sure where you're getting at with the >$200 restriction though. Unless it's that a lot of folks have already gone out and got far more expensive toys than that to do as little as that with. But something like $200 instead of $1400 is definitely lowering the boundary between the information-haves and info-have-nots, which is probably a good thing.
> Domain names are the addresses computer users enter to access a particular Web site. They come between "www" and "com," "net" or "org."
:]
Aaaaaargh. They come before ".us", don't they?
> the Clinton administration opposes the legislation, arguing that the best place to resolve domain-name disputes is in the courtroom.
Like how are they arguing? I see no links, no discussion... I don't think taking everything to court is a particularly productive way of settling every little dispute under the sun, although it might be considerably better than domain-names being bought-out for ever-escalating sums. Can't they just agree to let some friendly InterNIC-derivative have absolute control over what domains an organization is allowed?
Looks like pretty lousy "reporting" (CNN? I thought they were going up in the world?) and some bad ideas..
>I just recently took a class in X-Windows. :)
;]
...etc.
Yes, well Motif isn't exactly the best way of doing everything, or so I've heard (comments about square roots and roman numerals come to mind
> As to your pity being directed at us poor VB programmers, thanks but no thanks. VB has an amazingly productive UI. It sure beats the hell out of 4 xterms and writing my own make files.
That's up to the individual, I think. But give glade a blast before writing the non-VB world off as "4 xterms and writing your own Makefiles"!
*Cough*. From experience, VB3 produced executables with only dependencies being vbrun*.dll.
... and now I'm working in the IT industry (surprisingly enough) I'm not doing that much actual coding. Perl scripting is prevalent, java does the rounds, and the mainstay is C.
Let's just say that when I did pre-University work experience, I learnt & hacked VB3 and above. When I got to Uni, I was doing C, C++, ML,
Actually there are two flip-sides to what you say. Some thoughts that come to mind:
:)
/. before now about this, I'm sure :)
:) and decided, amidst the insults, to have a look at his VB project. After looking through it for 15minutes, despite his bad attitude I was thinking "this chap is paid, what, £25k/pa, to spend all day every day looking at this interface hacking this language?!", and I actually felt sorry for the chap!
:)
It's all very well everything being "consistent" but if you're consistently Microsoft, you're consistently closed-source.
There comes a point where the warm fuzzies of "building" on existing APIs freeze their bits off and you realise you're just clicking-in more bloatware with tie-ins to Windoze, to Micro$loth, to your particular development environment ("what vbrun300.dll?").
Does VB do containers for its controls like java, or like gtk?
Is COM really common? Is it free? Where are the implementations on Unix boxes? Do people even want such implementations?
You have a *major* misconception of how things operate in your phrase "Linux can't agree on the guts behind their GUIs", as though "Linux" were a company deserving of "their". To cut to the quick, don't do it
There is always glade, an open-source development environment for gtk+ pushing out C, C++ or Ada code (Perl to follow, at the last count). That's quite close to a "visual" environment, if you think such things are cool. That and I suspect you've never done any coding in xemacs, either.
Also you've got gnome/kde the wrong way round, IIRC: KDE is the one that uses some-CORBA,some-proprietary stuff; AFAIK gnome tries to do *everything* through CORBA. (There have been discussions on
It's fair enough that 'simple doesn't mean smart people shouldn't use it', but it does make it quite likely that there will be people out here who don't want to use it, quite often because it doesn't do all they want. Developers are and should be prepared to work to get things right.
A micro-story: at my last company, we had one customer from hell, as far as support was concerned. He was writing stuff in VB, using ODBC to connect, and he had a problem. I long-since decided it was a bug in VB, for various reasons (it was
Just my DM0,6-
You're right.
I, for one, never need go to http://www.lycos.co.uk/ (where the blasted thing redirects me to - talk about wasting bandwidth with DNS lookups), simply because there's nothing of any interest on there.
OTOH, I do prefer to use FTPSearch and MP3search, both of which are not expressly linked off the front page.
To me, it says it all that the actual web-searching bit has been reduced to a tiny box at the top of the page.
On the positive side, I noticed the other day that it seems to have turned into a bit of a portal: the question is, too little, too late?
> Of course, this is not flamebait - I don't necessarily hold the same opinion for animals which have been eradicated by humans.
Out of interest, why not?
OK, so we're the only species we know of to have evolved slashdot, but that's no reason to consider humans separately, especially *if* you think we 'evolved' from some other kind of mammal.
> Yes, it was obviously a joke -- however, I'm getting a little tired of the "it's redhat 4.x,5.x,6.x so it must be insecure" attitude paired with the "I've seen that movie hackers so it must be easy to hack" jokes.
I hope you're not suggesting that either attitude was present in my comment...
My point is that (a) a proper ruleset is always a good idea, (b) firewalls are still susceptible to breaches, be it by brute force or some other means (c) information given out makes folks' jobs easier to hack whatever it is, whether they should be able to get to port 23 or 80 or not.
<cynical>Oh look! A Unix workstation! I know how to use those!</cynical>
Then you, sir, are a scumbag and not a gentleman ;)
:)
Yes, it's blindingly obvious - it doesn't just have to www.royal.gov.uk, it could equally well be freshmeat.net - there are folks out & abouts who *don't* want it defaced. And if they're the ones providing the machinery to folks, they have the right to protect it. And letting on "it's RedHat 5.2" is probably supplying someone with a little too much. IMO
You must be aware you're looking for Debian, with its .deb format and particularly, apt utilities...
In particular, this link explains the differences fairly well. Being able to set up a list of sites with unstable and stable distributions and contrb / main / binary / non-free / non-US subdirectories, combine a list of package versions from them all, and then whenever you want to install something, 'apt-get install package' will do the job - INCLUDING dependencies.
At least RPM 3 has the ability to cope with HTTP URLs on the commandline - even if you do have to paste them in by hand....
Anyone know if it'll still have the Unique ID bug in it?
For some strange reason, I refuse to buy Intel at the moment...