Anybody working at SGI (Silicon Graphics before the big rebranding) will find this story extremely discouraging. It isn't just that it talks about Linux/commodity systems taking over a market that used to be dominated by Irix/SGI systems. It's the attitude expressed in the article, which has to have been picked up from the people the reporter interviewed: SGI products are legacy technology. They don't compete with Linux based systems -- they're just something people have to hold onto unti l the corresponding Linux technology matures enough for to replace it.
The one SGI product mentioned that is moving forward is the Linux version of Maya -- and the reporter seems unaware that Maya is an SGI product! A sad outcome for a company that once dominated computer animation.
I don't understand why people still go to headhunters. Even if this mysterious honest recruitment firm did exist, they'd still be taking a lot of money that could be going to your salary (they have to make a profit somehow), and they'll always be bad at matching you up with a company, because if they knew what they were talking about, they'd have a real job.
So everybody who knows anything about development does development? Good thing we never need to hire tech writers, managers... No wait, we do need these people. And people who are good at these jobs (and I agree, there are many incompetants) have a certain capacity to understand the technology. Not well enough to develop products, but well enough to document the work or manage or recruit the people who do the work.
Of course, many (perhaps most) tech writers, managers, and head hunters are technologically clueless. I hope you're not going to argue that this is no big thing. Maybe head hunters are superfluous, but you're naive if you think that we can dispense with tech writers or managers. Though many developers seem to think you can...
I talked to a head hunter a few days ago for a contract job. (You can't get contract work without dealing with recruiters, unless you have a lot of personal contacts.) At first I was appalled by her technical ignorance. But as she listened to my explanations of what I'd been doing, and how it might relate to the job she was trying to fill, I realized she had the native intelligence to do her job properly, even if it meant filling in her knowledge gaps on the fly. She'll do well. She is, alas, an exception.
What a good, honest head hunter brings to the table is connections. He or she has a relationship with a lot of managers, and has studied their needs carefully. The last time I got a permanent job through a head hunter, she earned her commission by connecting me with an obscure startup I never would have found on my own. And she did a better job of pitching me to the company than I ever could have done on my own. I simply would not have gotten the job without her help.
It's worth mentioning that one of the other head hunters in her firm tried to steal her commission. He got hold of my contact info, called me up behind her back, and persuaded me to go to an interview with one of his clients. As you might expect with such an unprincipled asshole, he hadn't made the slightest attempt to confirm that I was anything like a match for the job, and the whole thing was a big waste of time. Unfortunately, this jerk is far more typical than his co-worker.
And more so now, then when this all happened. This was before there were sites like monster.com, that help you make your own connections. Which seems to have discouraged the more honest and capable recruiters, but not the nitwits who just spam managers with poorly chosen resumes.
I'm not a graphic designer either -- like you I'm artistically challenged. But I certainly intend to write a style sheet or two for cZG. I might not submit it to them, and even if I did they'd probably reject it. But it's still a useful learning exercise. I have in mind something that's just as ugly as possible...
You raise a good point when you talk about designs that are not "robust against font size changes". You might find it productive to take one of these designs and fix it. I'm sure the original designer would find your changes instructive. In any case, it would be a very instructive exercise -- one I might try myself.
First off, note that all the designs on this site are user-submitted. If you think you can improve on them, please do so.
Second, this isn't a "cool web pages" site. Or perhaps I should say it isn't just a CWP site. It's main goal is to promote better web design techniques. That means limiting yourself to non-deprecated HTML and to the subset of CSS that is actually supported by most browsers. This makes for cleaner, more accessible, and more maintainable web pages. But it also limits what you can do in terms of fancy layout and formatting techniques.
Making sure all your text is text is not a "trick". It's just a design style that makes web pages 508 compliant. The only reason you've never seen it before is that most web designers (and most people who train them or write books for them) don't know or care about standard compliance.
I do admire the css Zen Garden, and also David Shea's other online design work. But not so much for the pretty graphics (nice, but not of extreme interest to an artistically challenged dweeb like me) as the way it promotes usable and accessible web content. Most web pages put a priority on presentation over content. Being a technical writer, I have exactly the opposite priority. Shea is doing a good job of promoting web design techniques that let you have it both ways.
I wonder if RMS thought he'd see a US presidential candidate releasing stuff under the GPL when he founded GNU 20 years ago!
RMS's goal was and is to get rid of "non-free" software, which he considers immoral. If he had any vision of 2002, it would have been of everybody using the GPL, or something similar.
Petty of me to pick this nit, but: RMS did not "found" GNU. He founded FSF. GNU is not an organization, it's an operating system, intended as an alterntive Unix. And, like his other grand plan, it is still unfinished!
Sure, if everybody does all their campaigning online. But as long as most campaigning consists of buying TV and radio time, running for office will be expensive.
Dean has done well so far by tapping online resources and communities. But remember that we haven't even started picking convention delegates yet. Once the primaries and caucuses start, Dean will have to find a way to get to all the voters and caucusers who aren't internet geeks. Maybe he can leverage his existing following into some kind of alternate campaign machine. But it's more likely that he'll just start spending money like any other candidate.
You need to cut back on the caffeine. At least I hope it's caffeine...
I stole the fee idea from Private Citizen. I first heard about them some years ago, when their founder got interviewed on NPR. At the time, he was just distributing information on how to sue telemarketers in Small Claims Court. (He described the attitude of his first trial judge to the defendent: "Last evening I had my dinner interrupted three times. You lose buddy.") Now I see that they offer a subscription where they publish your name on their own do-not-call list and sue violators for you.
It looks great in IE too. But only because the site has this policy:
We would like to see as much CSS1 as possible. CSS2 should be limited to widely-supported elements only. The css Zen Garden is about functional, practical CSS and not the latest bleeding-edge tricks viewable by 2% of the browsing public.
It's pretty sad that CSS2, which was published five years ago, is still considered "bleeding edge". Unfortunately, they're correct: nobody seems interested in a comprehensive implementation of CSS2. Which somehow hasn't prevented the W3C people from spending a lot of time on CSS3...
When you see a fancy CSS-based web site, it's important to remember that you don't need CSS to create a pretty web page. In the current state of the art, it's actually easier to create fancy eye candy using legacy HTML. What CSS buys you is a web page that's much easier to design and maintain, and is much less likely to have browser-specific issues. It's particularly important that a well-designed CSS-based page is still usable on a browser that doesn't even support CSS. It's not as pretty, but at least all the content is readable and organized the way the author meant it to be.
So you think that extortion is not a valid business model? The Mafia has made it work for years. Of course, the Mafia is something that always operates "somewhere else", but we all know people who would provide financial backing to your neighborhoold bully if they thought they could get away with it. Venture capitalism has never been known for strong social ethics.
You're right, most web pages are too browser-specific. (Irony of ironies, slashdot.org has serious display issues on Netscape 7.1.) But it's not because web developers are lazy or stupid. It's because of a very basic mistake made by the people who invented the web.
The original idea behind the web was that it was a simple distributed application for sharing information at places like CERN. They didn't worry about look-and-feel issues because they didn't think there would be any. If you're just using the web to share your informal notes on your latest particle physics experiment, you don't care exactly what your document elements look like, as long as the overall organization is clear to the reader. Thus the very first HTML specification says things like:
A list is a sequence of paragraphs, each of which is preceded by a special mark or sequence number.... The representation of the list is not defined here, but a bulleted list for unordered lists, and a sequence of numbered paragraphs for an ordered list would be quite appropriate.
HTML was also designed to make web pages easy to author. Which meant that web browsers had to be tolerant of small mistakes. Both these principles forced browser implementers to make a lot of choices about presentation without any guidance from the standards.
What the web's inventors didn't anticipate is that the web would become a mass medium. (So much so that most web users don't even grasp that the web and the internet are not the same thing.) When they saw this happening, they starting thinking about presentation issues and inventing style-sheet languages. But it took them years to thrash out new HTML and CSS specifications that dealt with these issues. But at the same time (1995 and thereabouts) Netscape was gearing up to deliver commerical web software. (Microsoft was still in denial about the Internet, and had launched MSN using proprietary client-server protocols.) Netscape wasn't willing to wait even a few months for W3C to deliver preliminary specifications: they saw their window of opportunity closing. So they invented their own. They took a lot of flack for this from people who pointed out that Netscape HTML violated all the rules of good markup. But Netscape wasn't interested in those issues -- they just wanted to deliver a browser that could display pretty web sites.
And of course when Microsoft joined the fray, they had the same attitude, only more so. So all the thousands of newly-minted web designers were indoctrinated to think in terms of tweaking proprietary HTML until it displayed correctly on one or two "preferred" browsers.
It's going to take a long time to reverse this mindset, and I doubt if it will ever disappear completely. It certainly doesn't help that no browser properly supports all the features of CSS2, which is the most important tool of the standard-complying web designer.
Taking that a step further: one popular anti-telemarketing startegy is to inform people that you're going to charge them a fee for dealing with their phone calls. Once they've been informed of this, they legally incur a financial obligation every time they call. This should be legally valid in Candada too.
Hmm, I can't document that Flemming ever actually refused to patent his discovery. Perhaps I'm guilty of spreading an urban legend. But the military funding of penicillin production research is well documented.
This is sort of ad hominem, but it's hard to accept criticism of markup languages from a guy whose web pages are hand-formatted text!
But let's skip the Latin and look at Hickson's actual arguments. They're a little convoluted, so maybe I'm not reading him correctly. As far as I can tell, he's mostly saying that correct XHTML is hard to prepare. Well, yeah, that's the whole point of using a tight-assed XML document type like XHTML instead of a tolerant, laid-back SGML document type like HTML: you're embracing restrictions that promote a well structured document. Following all the XML rules in XHTML is hard, and it's supposed to be. That's why I never prepare an XHTML document by hand. It's not that I don't believe in the goals of XHTML (structure, separating content and presentation), I just don't find it practical to manually touch all the bases that an XML document type requires me to touch. I use HTML 4, and avoid deprecated legacy tags and attributes. If you ever need a proper XML document, it's not that hard to do the conversion, probably using some kind of standard utility that will remember all the little details I'd forget.
There are souls brave enough to prepare XHTML or other XML documents by hand. That's not a mistake in and of itself, but it's foolish to do so without passing the document through a validator. Actually, the best XML editors validate your document on the fly.
Hickson also criticizes XHTML for its inconsistent browser support. Well, that's an issue with HTML as well, and something we need to take up with the browser vendors, not the XHTML advocates.
Hickson makes the assumption (one I once shared) that XHTML is primarily for delivering web content. As I said in a previous post, that's not correct.
For all its faults, the FBI is a lot less obnoxious than Britain's MI5. The FBI has to at least pretend to operate openly, and can get in trouble if they're caught ignoring people's rights. MI5 operates in secret, and can suspend people's right by administrative fiat.
Well first of all, legacy HTML will never go away -- not as long as millions of people are hacking out web pages by hand, or using antiquated HTML editors. XHTML will never completely replace legacy HTML, and if I still thought that was XHTML's central purpose, I would still consider XHTML a waste of effort.
The big virtue of XHTML is the big virtue of all XML document types: it's open. You can do anything with an XML document. I suppose that's also true of say TeX or RTF. Except these formats are very messy, and it's hard to extract the content from them. A good XML document type is well-structured, and thus relatively easy to access and manipulate.
If all you want to do with a document is display it as a single web page, that's not a big deal. But suppose you want to add it to some well-structured document management system? Or make it a chapter in a book? Or deliver it to a cell phone browser that uses WML or some other simplified markup language? Then all you have to do is write a filter that transforms your XHTML into the necessary XML document type. The possibilities are endless, and all of them are enabled by the simple openness of XML.
There are pitfalls, of course. A good XML application is carefully structured, and thoroughly separates presentation (layout, fonts, etc.) from content. That's why XHTML deprecates the use of formatting tags, like <center> and <font>, which act as if they designate content, but actually designate presentation. But there's nothing to prevent XHTML users from using deprecated features, or designers of other XML applications from structuring their documenting carelessly. So even after you run your document through HTML Tidy, you still might have to jump through a few hoops to transform it into a more sophisticated XML document type, such as DocBook. But the openness of XML makes just hoop-jumping a lot easier.
Anybody who's interesting in playing the XML transformation game needs to learn to program in the #1 XML transformation language, XSLT. This person has written some good introductory material, both online and in book form. Plus her web site neatly demonstrates the flexibility of the technology she teaches and advocates.
I've sneered at XHTML in the past, but I was speaking out of ignorance. I was assuming it was just a silly attempt to preserve HTML in an XML world. Actually, it's a very convenient bridge between HTML and XML. It's only incidentally about web content, since browsers will always need to support legacy HTML, and thus will never adopt all of XHTML's structure and restrictions. But once you have your content in XHTML format, you can transform it into any XML application you choose, using XSLT scripts. Which opens up a whole world of possibilities for people with all their content in messy old word processor formats, since word processor now tend to come with HTML export filters.
And there are other telnet clients besides Microsoft's. You can even use an X terminal window if you want to go to the trouble of installing Cygwin. Still, Drey has a point about those other MUD client features.
Slashdot Mirror
The one SGI product mentioned that is moving forward is the Linux version of Maya -- and the reporter seems unaware that Maya is an SGI product! A sad outcome for a company that once dominated computer animation.
Of course, many (perhaps most) tech writers, managers, and head hunters are technologically clueless. I hope you're not going to argue that this is no big thing. Maybe head hunters are superfluous, but you're naive if you think that we can dispense with tech writers or managers. Though many developers seem to think you can...
I talked to a head hunter a few days ago for a contract job. (You can't get contract work without dealing with recruiters, unless you have a lot of personal contacts.) At first I was appalled by her technical ignorance. But as she listened to my explanations of what I'd been doing, and how it might relate to the job she was trying to fill, I realized she had the native intelligence to do her job properly, even if it meant filling in her knowledge gaps on the fly. She'll do well. She is, alas, an exception.
What a good, honest head hunter brings to the table is connections. He or she has a relationship with a lot of managers, and has studied their needs carefully. The last time I got a permanent job through a head hunter, she earned her commission by connecting me with an obscure startup I never would have found on my own. And she did a better job of pitching me to the company than I ever could have done on my own. I simply would not have gotten the job without her help.
It's worth mentioning that one of the other head hunters in her firm tried to steal her commission. He got hold of my contact info, called me up behind her back, and persuaded me to go to an interview with one of his clients. As you might expect with such an unprincipled asshole, he hadn't made the slightest attempt to confirm that I was anything like a match for the job, and the whole thing was a big waste of time. Unfortunately, this jerk is far more typical than his co-worker.
And more so now, then when this all happened. This was before there were sites like monster.com, that help you make your own connections. Which seems to have discouraged the more honest and capable recruiters, but not the nitwits who just spam managers with poorly chosen resumes.
You raise a good point when you talk about designs that are not "robust against font size changes". You might find it productive to take one of these designs and fix it. I'm sure the original designer would find your changes instructive. In any case, it would be a very instructive exercise -- one I might try myself.
Second, this isn't a "cool web pages" site. Or perhaps I should say it isn't just a CWP site. It's main goal is to promote better web design techniques. That means limiting yourself to non-deprecated HTML and to the subset of CSS that is actually supported by most browsers. This makes for cleaner, more accessible, and more maintainable web pages. But it also limits what you can do in terms of fancy layout and formatting techniques.
I completely agree. Did you mean to respond to this post? Or maybe to the main story? Slashdot is confusing that way.
I do admire the css Zen Garden, and also David Shea's other online design work. But not so much for the pretty graphics (nice, but not of extreme interest to an artistically challenged dweeb like me) as the way it promotes usable and accessible web content. Most web pages put a priority on presentation over content. Being a technical writer, I have exactly the opposite priority. Shea is doing a good job of promoting web design techniques that let you have it both ways.
I shouldn't complain at being modded up, but WTF is this "informative"?
We're talking free software here. SecureCRT costs $99!
Petty of me to pick this nit, but: RMS did not "found" GNU. He founded FSF. GNU is not an organization, it's an operating system, intended as an alterntive Unix. And, like his other grand plan, it is still unfinished!
Dean has done well so far by tapping online resources and communities. But remember that we haven't even started picking convention delegates yet. Once the primaries and caucuses start, Dean will have to find a way to get to all the voters and caucusers who aren't internet geeks. Maybe he can leverage his existing following into some kind of alternate campaign machine. But it's more likely that he'll just start spending money like any other candidate.
I stole the fee idea from Private Citizen. I first heard about them some years ago, when their founder got interviewed on NPR. At the time, he was just distributing information on how to sue telemarketers in Small Claims Court. (He described the attitude of his first trial judge to the defendent: "Last evening I had my dinner interrupted three times. You lose buddy.") Now I see that they offer a subscription where they publish your name on their own do-not-call list and sue violators for you.
When you see a fancy CSS-based web site, it's important to remember that you don't need CSS to create a pretty web page. In the current state of the art, it's actually easier to create fancy eye candy using legacy HTML. What CSS buys you is a web page that's much easier to design and maintain, and is much less likely to have browser-specific issues. It's particularly important that a well-designed CSS-based page is still usable on a browser that doesn't even support CSS. It's not as pretty, but at least all the content is readable and organized the way the author meant it to be.
So you think that extortion is not a valid business model? The Mafia has made it work for years. Of course, the Mafia is something that always operates "somewhere else", but we all know people who would provide financial backing to your neighborhoold bully if they thought they could get away with it. Venture capitalism has never been known for strong social ethics.
The original idea behind the web was that it was a simple distributed application for sharing information at places like CERN. They didn't worry about look-and-feel issues because they didn't think there would be any. If you're just using the web to share your informal notes on your latest particle physics experiment, you don't care exactly what your document elements look like, as long as the overall organization is clear to the reader. Thus the very first HTML specification says things like:
HTML was also designed to make web pages easy to author. Which meant that web browsers had to be tolerant of small mistakes. Both these principles forced browser implementers to make a lot of choices about presentation without any guidance from the standards.
What the web's inventors didn't anticipate is that the web would become a mass medium. (So much so that most web users don't even grasp that the web and the internet are not the same thing.) When they saw this happening, they starting thinking about presentation issues and inventing style-sheet languages. But it took them years to thrash out new HTML and CSS specifications that dealt with these issues. But at the same time (1995 and thereabouts) Netscape was gearing up to deliver commerical web software. (Microsoft was still in denial about the Internet, and had launched MSN using proprietary client-server protocols.) Netscape wasn't willing to wait even a few months for W3C to deliver preliminary specifications: they saw their window of opportunity closing. So they invented their own. They took a lot of flack for this from people who pointed out that Netscape HTML violated all the rules of good markup. But Netscape wasn't interested in those issues -- they just wanted to deliver a browser that could display pretty web sites.
And of course when Microsoft joined the fray, they had the same attitude, only more so. So all the thousands of newly-minted web designers were indoctrinated to think in terms of tweaking proprietary HTML until it displayed correctly on one or two "preferred" browsers.
It's going to take a long time to reverse this mindset, and I doubt if it will ever disappear completely. It certainly doesn't help that no browser properly supports all the features of CSS2, which is the most important tool of the standard-complying web designer.
Taking that a step further: one popular anti-telemarketing startegy is to inform people that you're going to charge them a fee for dealing with their phone calls. Once they've been informed of this, they legally incur a financial obligation every time they call. This should be legally valid in Candada too.
Hmm, I can't document that Flemming ever actually refused to patent his discovery. Perhaps I'm guilty of spreading an urban legend. But the military funding of penicillin production research is well documented.
But let's skip the Latin and look at Hickson's actual arguments. They're a little convoluted, so maybe I'm not reading him correctly. As far as I can tell, he's mostly saying that correct XHTML is hard to prepare. Well, yeah, that's the whole point of using a tight-assed XML document type like XHTML instead of a tolerant, laid-back SGML document type like HTML: you're embracing restrictions that promote a well structured document. Following all the XML rules in XHTML is hard, and it's supposed to be. That's why I never prepare an XHTML document by hand. It's not that I don't believe in the goals of XHTML (structure, separating content and presentation), I just don't find it practical to manually touch all the bases that an XML document type requires me to touch. I use HTML 4, and avoid deprecated legacy tags and attributes. If you ever need a proper XML document, it's not that hard to do the conversion, probably using some kind of standard utility that will remember all the little details I'd forget.
There are souls brave enough to prepare XHTML or other XML documents by hand. That's not a mistake in and of itself, but it's foolish to do so without passing the document through a validator. Actually, the best XML editors validate your document on the fly.
Hickson also criticizes XHTML for its inconsistent browser support. Well, that's an issue with HTML as well, and something we need to take up with the browser vendors, not the XHTML advocates.
Hickson makes the assumption (one I once shared) that XHTML is primarily for delivering web content. As I said in a previous post, that's not correct.
For all its faults, the FBI is a lot less obnoxious than Britain's MI5. The FBI has to at least pretend to operate openly, and can get in trouble if they're caught ignoring people's rights. MI5 operates in secret, and can suspend people's right by administrative fiat.
Though the markup on markup can be pretty horrendous!
The big virtue of XHTML is the big virtue of all XML document types: it's open. You can do anything with an XML document. I suppose that's also true of say TeX or RTF. Except these formats are very messy, and it's hard to extract the content from them. A good XML document type is well-structured, and thus relatively easy to access and manipulate.
If all you want to do with a document is display it as a single web page, that's not a big deal. But suppose you want to add it to some well-structured document management system? Or make it a chapter in a book? Or deliver it to a cell phone browser that uses WML or some other simplified markup language? Then all you have to do is write a filter that transforms your XHTML into the necessary XML document type. The possibilities are endless, and all of them are enabled by the simple openness of XML.
There are pitfalls, of course. A good XML application is carefully structured, and thoroughly separates presentation (layout, fonts, etc.) from content. That's why XHTML deprecates the use of formatting tags, like <center> and <font>, which act as if they designate content, but actually designate presentation. But there's nothing to prevent XHTML users from using deprecated features, or designers of other XML applications from structuring their documenting carelessly. So even after you run your document through HTML Tidy, you still might have to jump through a few hoops to transform it into a more sophisticated XML document type, such as DocBook. But the openness of XML makes just hoop-jumping a lot easier.
Anybody who's interesting in playing the XML transformation game needs to learn to program in the #1 XML transformation language, XSLT. This person has written some good introductory material, both online and in book form. Plus her web site neatly demonstrates the flexibility of the technology she teaches and advocates.
That's not saying much!
I've sneered at XHTML in the past, but I was speaking out of ignorance. I was assuming it was just a silly attempt to preserve HTML in an XML world. Actually, it's a very convenient bridge between HTML and XML. It's only incidentally about web content, since browsers will always need to support legacy HTML, and thus will never adopt all of XHTML's structure and restrictions. But once you have your content in XHTML format, you can transform it into any XML application you choose, using XSLT scripts. Which opens up a whole world of possibilities for people with all their content in messy old word processor formats, since word processor now tend to come with HTML export filters.
And (to bring us back ontopic), TinyFugue runs under Windows using Cygwin.
I've played MUDs. I have to admit I'm not very good at RPGs, which is probably why I never graducated beyond Telnet mode.
And there are other telnet clients besides Microsoft's. You can even use an X terminal window if you want to go to the trouble of installing Cygwin. Still, Drey has a point about those other MUD client features.