You can always tell a M$ designed webpage because it has a grey background (the default in IE is white so they don't bother to set it).
Gosh, I never knew the webpages I make using Linux were M$ designed. Believing that the reader knows best what his or her preferred background is, is an insight shared by more people than the ones on M$'s payroll.
Then if the Disney corporation registers disney.per, an individual with the last name of Disney should be able to take them to court and sue them for wrongfully and maliciously registering a domain name that they're not entitled to.
But what if Walt still had been alive? And besides, there's probably more than one person named Disney. Can Joe Disney sue Mary Disney over the right to use 'disney.per'? What's going to happen these new TLDs come into existence, and withing the hour, one million Koreans claim 'kim.per'?
I expect that any telnet client that allows telnetting to such a domain name is not implementing the standards for domain names correctly
I wouldn't expect any telnet or ftp client to do domainname lookups themselves. That's why we have gethostbyname(3), don't we?
However, typical RFC standards recommend "be conservative in what you send, be liberal in what you accept". I very much doubt an RFC says a client MUST NOT be able to connect to badly formatted hostname. It probably doesn't even say the client SHOULD NOT be able to connect.
Is there any good reason any more to not allow the registration of infinite TLD's? ... Large websites could be organized by TLD instead of directory. IBM could have WWW.IBM.SALES for sales, and WWW.IBM.SUPPORT for help, and WWW.IBM.NYSE for investor information.
That wouldn't solve the problem at all, would it? That would mean "IBM" owns all of domains IBM.TLD, for all possible TLDs. Yes, you have created more space, but not where it matters: the part in front of the TLD. Not to mention the headaches of 'foo.com', 'foo.net', 'foo.org' and all the 'foo.xx's in various countries sueing each other over the rights of using 'foo.sales' and 'foo.support'.
But even creating an infinite number of TLDs, and have different entities take 'foo.shop', 'foo.store', 'foo.buy', etc will not solve the problem. Look at etoy vs etoys. The claim of etoys is the confusion of a similar name. You'd think that they let someone else take etoys.buy, etoys.toy or etoys.comm if they can prevent it?
The current problem with the DNS name space isn't a technical one - it's a social one. A technical solution of just adding more TLDs isn't going to solve the problem.
It appears to me that the percentage of women in the computer industry (in non-clerical or management positions) has increased significantly over the last decade.
Unfortunally, my experience is quite the opposite. I started my life in the "computer world" about 20 years ago, when I started studying computer science and math. My first 10 years were mainly academic. While a minority, female CS or math scientists are not uncommon. Female (full) CS professors, women chairing important conferences, all common enough not have warranted an article like the one discussed now, had the article been about the (academic) CS community 10 years ago. I'd say 20-25% is female.
But then, after the academic world, I moved to other areas. Worked for an ISP, I made web sites, then moved to US, were I did tech support, worked as a tech analyst, and later as a database administrator. I noticed that the non-academic tech world certainly had a lot less women academic world. And, while in the academic world, the male/female ratio seems to be the same from top to bottom (professors, researchers, students), that's not the case in the non-academic world. Women are under represented in the higher positions.
But what's even more shocking is the Perl world - a subset of the computer world. There are hardly any female Perl people. I was at YAPC last year, a Perl conference. 2 talks by women, and about 5% of the attendees were female. There are no women who are important Perl porters, hardly any female CPAN authors. There are a few women on #perl, the IRC channel, but that's still a tiny minority.
I don't know why women are a minority in the computer world, and I don't know why it's even worse in the Perl community. And unfortunally, I don't see any tendency of it getting better.
However, it is quite sad that not many women decided to follow in her footsteps, because I think it would be really cool to date a geek girl.
Urg. That attitude is one of the reasons women stay away from the computing world. Women in the computer world are a minority, and the prospect of a whole bunch of men whose first thought is "it would be cool to date her" is enough to keep groups of women away.
I futher would like to point out that geek != tech. I know a lot of tech (wo)men who certainly aren't geeks and I know a lot of geek (wo)men who I wouldn't classify as tech savvy.
The first computer coders were women. Check your history of computers: they were initially coded by connecting wires in a switchboard configuration. Who do you think did that work?
I would even go further, and say two of the people who influenced current programming the most were women. Ada Lovelace, who studied and wrote about computer programming, long before there were working computers, and Grace Hopper who pioneered the concept of programming languages and compilers.
Unfortunally, I cannot name any woman after Grace Hopper.
Everyone on the Internet is gender neutral until I meet them in meatspace.
Perhaps for you, but not for everyone. 15 years on "the net" have resulted in getting contacted by a long list of people who are only interested in you because you have a female name.
There are quite a few women out there who use male sounding names to prevent them from being treated as a stupid person, or a lust object.
If we go by that reasoning, then Slashdot would be better off as a newsgroup, as web sites aren't the ideal tool to do public messages.
Oh, most certainly. Message boards like Slashdot are by far the most crappiest reinvention of usenet I can imagine. Not even Microsoft has ever pulled a stunt like that. Of course, Usenet doesn't come with ads, so I guess Slashdot will never turn into a newsgroup. But just think of the possibilities: usenet clients keep track of which articles you've read, so you can go back to a story, and see which comments you haven't read. You can use your own score files - don't like a person? Never see her comments again. You would be able to not see any AC post, but all signed posts, regardless of their moderation. Still want to use the current moderation? NoCem will do that. And in good Unix tradition, you would be able to use your own editor instead of a crappy web form. Not to mention to automatically archive your posts, if you want to, or send an email to the poster instead of posting a followup. Ah, good old Usenet.
And now a days, there are even point-and-drool clients, build in into your web wowser.
i like being able to post as AC. but if no one will ever see it, than what's the point?
The only AC posts I see are those that have been moderated up. Even if I have moderation points. Would slashdot have killfiles, like Usenet clients, I would kill all AC posts. There are a few exceptions, but in general, if you don't find your article worthy enough to put your name on it, why should I find it worthy enough to read it?
...the first fifty get much more moderation than the last 150 (both positive and negative). It would be good if moderators could be given a random list of say 20 comments with the suggestion that they read this list first.
That of course would not solve the problem. The most important reason the first 50 get moderated more is that the first 50 have been there longer. If a moderator comes in, and there are only 30 posts yet, all (s)he can do is moderate those 30 posts. And unlike even the crappiest newsreader, slashdot doesn't keep track which comments you have already seen, so it really discourages someone from visiting a thread for a second time.
You want a useful moderation scheme? Use Usenet and a NoCem client.
We've decided to squash the bugs and make a clean release rather than rush it.
Isn't this one of the MAIN reasons Open Source is supposed to be so good?
NO!
The last thing the Open Source movement needs is give the impression that Open Source products are buggy, and you have to fix your own bugs. One of the main problems people have in using Open Source products in their believe the quality isn't good.
Ah, back in the good old days, people knew what an email address was for. And in good old Unix tradition, they used one tool for one job. For email, they'd use/bin/mail or one of its friends.
But Redmond style "lets bloat everything so everything has everything buildin" has taken over websites as well. After reinventing usenet (in a poor way), slashdot is now going to reinvent email as well!
If I want to send a private message to you, I'd email douglas@min.net. But I guess that would be lost to you, as you seem to think web sites are the ideal tool to do private messages.
It's progress I guess. It must be cool for spammers though... targetted audience! Oh, wait. I suddenly realize the big advantage. A private message system on slashdot means more banner ads to sell! You won't get banner ads with email!
Am I missing something? Can't they just ignore the patches and the changes?
A lot of people are screaming because they get a whopping 5 spam messages a week. How much requests for help do you think CmdrTaco and Hemos would get, even if they don't respond to them?
But there's more. Look for instance at the infamous Matt Wright. He made a whole bunch of free CGI programs. They aren't supported. So, even if patches are made, they aren't merged back in. There are problems with the programs. The Usenet group comp.lang.perl.misc, a group that doesn't deal with CGI, and where Matt Wright doesn't post, gets several messages *daily* regarding his programs. Even now, for programs that were written years ago. His programs might have helped a few, but they caused headaches for many people - including people that not even use the programs. And it didn't do much good for Matt's name either.
I think people who don't want release their code before it's finished shouldn't. And I don't think the Open Source community needs people with a gimme, gimme, gimme attitude.
We're just too busy to play tech support helping dozens of people compile mod_perl and tune Apache. We've decided to squash the bugs and make a clean release rather than rush it.
This is the attitude that makes people write you flames about not turning your code loose. You aren't the only people who can compile mod_perl and support Apache. And, believe it or not, there are people out here who are intelligent enough to understand your code without you having to explain it.
He didn't say noone could. But for anyone who can, and takes the time to figure it out, there will be hundreds of requests for help.
Release the code as extreme alpha, refuse to support Apache and mod_perl, and put your money where your mouth is about being a member of the OpenSource community.
Eh, who says that OpenSource means "release something that isn't finished"? Is there someone like RMS, ESR or Bruce Perens claiming that you have to release unfinished products? I don't think so.
Your attitude belongs more to the gimme, gimme, gimme warez-kiddies than it belongs to the Open Source movement.
And another device that would act as a becon, in case we "lose" another probe, so we could at least locate it and see if there is anything salvagable from the crash site.
Right.... and who's going there and pick it up? There's no point in exactly knowing where a piece of junk on Mars is. It no longer works, and hence, it's useless. Sending something there with equipment to "salvage" things from the crash and then bring it back is much, much more expensive than just to rebuild it. It's faster to rebuild as well, as a round trip takes quite some time.
AI systems which could simulate the emergency decision making processes of a crack pilot like Armstrong.
It doesn't take AI or a crack pilot to trigger Big rock! Avoid!. It does however take quite some computing power to recognize a big rock while descending from orbit to the surface.
I'd aim for Kansas, or Florida, or the Sahara (flat, flat, not flat but reasonably soft).
Earth has a lot of places that are much flatter than Mars. A thicker atmosphere, weather, and most of all, the presence of water make for erosion and the settling of sediments that make flat areas.
I do however disagree with the Sahara being reasonble soft. I've been there, and the Sahara isn't made of the same sands as a beach. It's very rocky, and I imagine it's a lot like the surface of the Moon or Mars, minus the craters and canyons.
The problem is that the public's imagination is captured by manned space flight.
Really? Isn't one of the reasons there were only a few missions to the moon that the publics interest was lost quickly?
Should we spend more than M$ advertisement department on a man mission to Mars, just as a hope to keep the public interested? Do you really think people in this era can have their attention grabbed for the year and a half or so it takes to send a tin can with 4 boring people to a big red rock? You wouldn't be able to sell a soap opera with that theme to a network, let alone do the real thing.
I don't think the general public is interested in a multi year mission to Mars and back.
This is *NOT* a perl problem, nor a design flaw, nor anything of the sort.
It is a Perl problem in the sense that it effects Perl too - it is however not only a Perl problem. Perhaps it isn't a design flaw, but it certainly isn't the smartest decision made by the people making the C standard. Returning years since 1900 doesn't help the programmer at all - now he needs to do work regardless whether he needs 2 digit years or 4 digit years. Years since 1 BC (the year before 1 AD, or the year 0 for the astronomers under us) would have avoided a lot of problems.
While not reading the manual isn't an excuse, you can't get away with stupid decisions just because you've documented it. I think it was a very unfortunate decision. It is of course way to late to turn it back. When designing Perl, Larry could have choosen that localtime would return 4 digit years. He preferred to follow the C standard, which has merrit as well. Had Perl be designed 10 years later, he might have made another decision.
First, the type system, or more precisely, the lack of one. Values are not type-safe, in the sense that they can change meaning based on the context they show up in (eg, strings and numbers).
Natural languages aren't context free, and people have surprisingly few problems with that. Why shouldn't a computer language use context?
When you want to change the type of a value you do it explicitly.
And doing it explicitly and not implicitly is good because of...?
In a language with well-defined types, it's trivial to overload operators so they do the right thing polymorphically
True, but look what you are saying later:
Syntax is helpful when it distinguishes different semantic domains;
So, first you bash Perl from not treating strings and scalars are different semantic domains, but yet is applaud the fact that it's possible to overload operators to do semantically different things based on the types of their operands. That doesn't make any sense. Could you please make up your mind which of the two you like?
Argument list flattening. Again, there are ways around this, but they require fairly sophisticated understanding of the language.
That's like saying that using pointers requires a fairly sophisticated understanding of C.
In fact, much of Perl's syntax is just pointless complexity.
That's not a fact, that's an opinion.
Why are references to aggregates prefixed with a '$'?
Because references are scalars and not aggregates. I don't find that hard to understand.
How come packages and classes are defined with the same syntax?
Because a class is a package. Remember: Syntax is helpful when it distinguishes different semantic domains. In Perl, there are no semantic differences between packages and classes.
Because Perl offers so much flexibility is it very difficult to understand code based on a knowledge of a subset of the Perl langauge
Can you point out a language where you can maintain code based on knowledge of only a subset of the language? Letting people maintain code who only know a subset of the language is a sure way of disaster - in any language.
because the syntax is so flexible, almost everybody is only aware of a subset of the language
but if you ever have to read code, flexibility and reduncancy(sp?) quickly become the enemy
This is true to a certain extend. That's why childrens books use a small set of words, and are often repeatative. But great novels tend to use a much larger set of words, and people don't mind different constructs of sentences.
Slashdot Mirror
Gosh, I never knew the webpages I make using Linux were M$ designed. Believing that the reader knows best what his or her preferred background is, is an insight shared by more people than the ones on M$'s payroll.
-- Abigail
But what if Walt still had been alive? And besides, there's probably more than one person named Disney. Can Joe Disney sue Mary Disney over the right to use 'disney.per'? What's going to happen these new TLDs come into existence, and withing the hour, one million Koreans claim 'kim.per'?
-- Abigail
I wouldn't expect any telnet or ftp client to do domainname lookups themselves. That's why we have gethostbyname(3), don't we?
However, typical RFC standards recommend "be conservative in what you send, be liberal in what you accept". I very much doubt an RFC says a client MUST NOT be able to connect to badly formatted hostname. It probably doesn't even say the client SHOULD NOT be able to connect.
-- Abigail
Large websites could be organized by TLD instead of directory. IBM could have WWW.IBM.SALES for sales, and WWW.IBM.SUPPORT for help, and WWW.IBM.NYSE for investor information.
That wouldn't solve the problem at all, would it? That would mean "IBM" owns all of domains IBM.TLD, for all possible TLDs. Yes, you have created more space, but not where it matters: the part in front of the TLD. Not to mention the headaches of 'foo.com', 'foo.net', 'foo.org' and all the 'foo.xx's in various countries sueing each other over the rights of using 'foo.sales' and 'foo.support'.
But even creating an infinite number of TLDs, and have different entities take 'foo.shop', 'foo.store', 'foo.buy', etc will not solve the problem. Look at etoy vs etoys. The claim of etoys is the confusion of a similar name. You'd think that they let someone else take etoys.buy, etoys.toy or etoys.comm if they can prevent it?
The current problem with the DNS name space isn't a technical one - it's a social one. A technical solution of just adding more TLDs isn't going to solve the problem.
-- Abigail
Most men don't either.
-- Abigail
Unfortunally, my experience is quite the opposite. I started my life in the "computer world" about 20 years ago, when I started studying computer science and math. My first 10 years were mainly academic. While a minority, female CS or math scientists are not uncommon. Female (full) CS professors, women chairing important conferences, all common enough not have warranted an article like the one discussed now, had the article been about the (academic) CS community 10 years ago. I'd say 20-25% is female.
But then, after the academic world, I moved to other areas. Worked for an ISP, I made web sites, then moved to US, were I did tech support, worked as a tech analyst, and later as a database administrator. I noticed that the non-academic tech world certainly had a lot less women academic world. And, while in the academic world, the male/female ratio seems to be the same from top to bottom (professors, researchers, students), that's not the case in the non-academic world. Women are under represented in the higher positions.
But what's even more shocking is the Perl world - a subset of the computer world. There are hardly any female Perl people. I was at YAPC last year, a Perl conference. 2 talks by women, and about 5% of the attendees were female. There are no women who are important Perl porters, hardly any female CPAN authors. There are a few women on #perl, the IRC channel, but that's still a tiny minority.
I don't know why women are a minority in the computer world, and I don't know why it's even worse in the Perl community. And unfortunally, I don't see any tendency of it getting better.
-- Abigail
Urg. That attitude is one of the reasons women stay away from the computing world. Women in the computer world are a minority, and the prospect of a whole bunch of men whose first thought is "it would be cool to date her" is enough to keep groups of women away.
I futher would like to point out that geek != tech. I know a lot of tech (wo)men who certainly aren't geeks and I know a lot of geek (wo)men who I wouldn't classify as tech savvy.
-- Abigail
I would even go further, and say two of the people who influenced current programming the most were women. Ada Lovelace, who studied and wrote about computer programming, long before there were working computers, and Grace Hopper who pioneered the concept of programming languages and compilers.
Unfortunally, I cannot name any woman after Grace Hopper.
-- Abigail
Perhaps for you, but not for everyone. 15 years on "the net" have resulted in getting contacted by a long list of people who are only interested in you because you have a female name.
There are quite a few women out there who use male sounding names to prevent them from being treated as a stupid person, or a lust object.
-- Abigail
Oh, most certainly. Message boards like Slashdot are by far the most crappiest reinvention of usenet I can imagine. Not even Microsoft has ever pulled a stunt like that. Of course, Usenet doesn't come with ads, so I guess Slashdot will never turn into a newsgroup. But just think of the possibilities: usenet clients keep track of which articles you've read, so you can go back to a story, and see which comments you haven't read. You can use your own score files - don't like a person? Never see her comments again. You would be able to not see any AC post, but all signed posts, regardless of their moderation. Still want to use the current moderation? NoCem will do that. And in good Unix tradition, you would be able to use your own editor instead of a crappy web form. Not to mention to automatically archive your posts, if you want to, or send an email to the poster instead of posting a followup. Ah, good old Usenet.
And now a days, there are even point-and-drool clients, build in into your web wowser.
-- Abigail
The only AC posts I see are those that have been moderated up. Even if I have moderation points. Would slashdot have killfiles, like Usenet clients, I would kill all AC posts. There are a few exceptions, but in general, if you don't find your article worthy enough to put your name on it, why should I find it worthy enough to read it?
-- Abigail
That of course would not solve the problem. The most important reason the first 50 get moderated more is that the first 50 have been there longer. If a moderator comes in, and there are only 30 posts yet, all (s)he can do is moderate those 30 posts. And unlike even the crappiest newsreader, slashdot doesn't keep track which comments you have already seen, so it really discourages someone from visiting a thread for a second time.
You want a useful moderation scheme? Use Usenet and a NoCem client.
-- Abigail
Isn't this one of the MAIN reasons Open Source is supposed to be so good?
NO!
The last thing the Open Source movement needs is give the impression that Open Source products are buggy, and you have to fix your own bugs. One of the main problems people have in using Open Source products in their believe the quality isn't good.
-- Abigail
Ah, back in the good old days, people knew what an email address was for. And in good old Unix tradition, they used one tool for one job. For email, they'd use /bin/mail or one of its friends.
But Redmond style "lets bloat everything so everything has everything buildin" has taken over websites as well. After reinventing usenet (in a poor way), slashdot is now going to reinvent email as well!
If I want to send a private message to you, I'd email douglas@min.net. But I guess that would be lost to you, as you seem to think web sites are the ideal tool to do private messages.
It's progress I guess. It must be cool for spammers though... targetted audience! Oh, wait. I suddenly realize the big advantage. A private message system on slashdot means more banner ads to sell! You won't get banner ads with email!
-- Abigail
A lot of people are screaming because they get a whopping 5 spam messages a week. How much requests for help do you think CmdrTaco and Hemos would get, even if they don't respond to them?
But there's more. Look for instance at the infamous Matt Wright. He made a whole bunch of free CGI programs. They aren't supported. So, even if patches are made, they aren't merged back in. There are problems with the programs. The Usenet group comp.lang.perl.misc, a group that doesn't deal with CGI, and where Matt Wright doesn't post, gets several messages *daily* regarding his programs. Even now, for programs that were written years ago. His programs might have helped a few, but they caused headaches for many people - including people that not even use the programs. And it didn't do much good for Matt's name either.
I think people who don't want release their code before it's finished shouldn't. And I don't think the Open Source community needs people with a gimme, gimme, gimme attitude.
-- Abigail
This is the attitude that makes people write you flames about not turning your code loose. You aren't the only people who can compile mod_perl and support Apache. And, believe it or not, there are people out here who are intelligent enough to understand your code without you having to explain it.
He didn't say noone could. But for anyone who can, and takes the time to figure it out, there will be hundreds of requests for help.
Release the code as extreme alpha, refuse to support Apache and mod_perl, and put your money where your mouth is about being a member of the OpenSource community.
Eh, who says that OpenSource means "release something that isn't finished"? Is there someone like RMS, ESR or Bruce Perens claiming that you have to release unfinished products? I don't think so.
Your attitude belongs more to the gimme, gimme, gimme warez-kiddies than it belongs to the Open Source movement.
-- Abigail
Right.... and who's going there and pick it up? There's no point in exactly knowing where a piece of junk on Mars is. It no longer works, and hence, it's useless. Sending something there with equipment to "salvage" things from the crash and then bring it back is much, much more expensive than just to rebuild it. It's faster to rebuild as well, as a round trip takes quite some time.
-- Abigail
It doesn't take AI or a crack pilot to trigger Big rock! Avoid!. It does however take quite some computing power to recognize a big rock while descending from orbit to the surface.
-- Abigail
Earth has a lot of places that are much flatter than Mars. A thicker atmosphere, weather, and most of all, the presence of water make for erosion and the settling of sediments that make flat areas.
I do however disagree with the Sahara being reasonble soft. I've been there, and the Sahara isn't made of the same sands as a beach. It's very rocky, and I imagine it's a lot like the surface of the Moon or Mars, minus the craters and canyons.
-- Abigail
Really? Isn't one of the reasons there were only a few missions to the moon that the publics interest was lost quickly?
Should we spend more than M$ advertisement department on a man mission to Mars, just as a hope to keep the public interested? Do you really think people in this era can have their attention grabbed for the year and a half or so it takes to send a tin can with 4 boring people to a big red rock? You wouldn't be able to sell a soap opera with that theme to a network, let alone do the real thing.
I don't think the general public is interested in a multi year mission to Mars and back.
-- Abigail
It is a Perl problem in the sense that it effects Perl too - it is however not only a Perl problem. Perhaps it isn't a design flaw, but it certainly isn't the smartest decision made by the people making the C standard. Returning years since 1900 doesn't help the programmer at all - now he needs to do work regardless whether he needs 2 digit years or 4 digit years. Years since 1 BC (the year before 1 AD, or the year 0 for the astronomers under us) would have avoided a lot of problems.
While not reading the manual isn't an excuse, you can't get away with stupid decisions just because you've documented it. I think it was a very unfortunate decision. It is of course way to late to turn it back. When designing Perl, Larry could have choosen that localtime would return 4 digit years. He preferred to follow the C standard, which has merrit as well. Had Perl be designed 10 years later, he might have made another decision.
-- Abigail
Now, if we only had a standard Java environment....
-- Abigail
Natural languages aren't context free, and people have surprisingly few problems with that. Why shouldn't a computer language use context?
When you want to change the type of a value you do it explicitly.
And doing it explicitly and not implicitly is good because of...?
In a language with well-defined types, it's trivial to overload operators so they do the right thing polymorphically
True, but look what you are saying later:
Syntax is helpful when it distinguishes different semantic domains;
So, first you bash Perl from not treating strings and scalars are different semantic domains, but yet is applaud the fact that it's possible to overload operators to do semantically different things based on the types of their operands. That doesn't make any sense. Could you please make up your mind which of the two you like?
Argument list flattening. Again, there are ways around this, but they require fairly sophisticated understanding of the language.
That's like saying that using pointers requires a fairly sophisticated understanding of C.
In fact, much of Perl's syntax is just pointless complexity.
That's not a fact, that's an opinion.
Why are references to aggregates prefixed with a '$'?
Because references are scalars and not aggregates. I don't find that hard to understand.
How come packages and classes are defined with the same syntax?
Because a class is a package. Remember: Syntax is helpful when it distinguishes different semantic domains. In Perl, there are no semantic differences between packages and classes.
-- Abigail
Can you point out a language where you can maintain code based on knowledge of only a subset of the language? Letting people maintain code who only know a subset of the language is a sure way of disaster - in any language.
because the syntax is so flexible, almost everybody is only aware of a subset of the language
And you back up your claim with... ?
-- Abigail
This is true to a certain extend. That's why childrens books use a small set of words, and are often repeatative. But great novels tend to use a much larger set of words, and people don't mind different constructs of sentences.
-- Abigail