There are currently thousands of highly qualified people available now who will work for half or even a third of the salary as average. This is why the government conspired with wall street to bust the big bubble, because no one would work for the government anymore (no stock options). And unlike H-1Bs, who have to be paid what the average person makes, you can legally pay Americans way under average. So now there are plenty of admins available... and programmers, too. Just post the openings here and watch the geeks resumes come flooding in.
There are lots of jobs like that in many companies. Instead of having one person do it for all the branches of various different companies in the same city (e.g. via a contracting firm that arranges it this way), they end up shuttling people all around. While one person flies from Dallas to Atlanta to install XP, another flies from Atlanta to Dallas to install XP. Well, I guess it helps the airlines; they need it.
There are good apples and bad apples in every bunch. I've met both good and bad H-1B people. I remember going on a sales visit (four of us, including our director of sales) to a major telco equipment manufacturer to demonstrate a software product. We were showing this to a couple of managers who were assisted by their lead system administrator (who ran a network of servers and workstations for an engineering group of 300 people). I believe there was one other sysadmin working there, but he was not in this meeting. While the sysadmin managed to install the package just fine, he wasn't able to configure the network on the machine (Solaris) to successfully talk to the clients. So I had to tell him what commands to type in. There were a few other cases of him not knowing what he was doing. And the two managers didn't seem to be surprised at this at all. Toward the end of the meeting I found out he was an H-1B worker from Egypt.
For every example someone can come up with for a good person, there will be another for a bad person. For every example someone can come up with for a bad person, there will be another for a good person. But what is a fact is that the US government simply does not do a thorough review of the H-1B applications. They don't even have staff qualified to understand the various sub fields that technology employers need. They just make sure the paperwork is right and the named person isn't a known terrorist and stamp "approved". The whole process for H-1B costs a fraction of what the payout to a technology talent recruiter would make placing just one really good techie.
During the boom, sure, it was hard to find people, but that wasn't from lack of people. It was from a non-streamlined recruiting process that in itself was very expensive, and doing a poor job of communicating and matching people (I was frequently called up for programming jobs, and that was not what my resume said I was looking for... either they can't read or they were extremely desperate). Even the job board web sites suck today. Put in "linux" as a keyword and you match lots of Windows jobs that say "experience with linux a plus".
YOUR company may not be abusing the system, but many companies are. One recent example exposed was Bank of America. Not only did they replace lots of IT workers with H-1B people, they even required the replaced domestic workers to stay a few weeks and train their H-1B replacements in order to qualify for the severance package. This is why I refuse to do business with BoA.
What makes you think that H-1B abuses get caught? The government isn't reviewing them. The companies doing the abuse certainly aren't telling. Yet people are being replaced by H-1B workers in both the boom and the bust.
It's the largest corporations that have the H-1B process streamlined where it's no longer a hassle for them.
And going back to school is not the answer. What would you do, get a 2nd CS degree to replace your first? If you go back to school I recommend getting a degree in Salesmanship... there is a current shortage of good sales people. What high tech businesses want in their employees is experience. Part of the problem where shortages exist is that there is less of a pathway to achieve experience than there has been before. Even during the boom, less experienced and inexperienced people could not find jobs (I know some personally who had this trouble). Another part of the problem is that as technology changes, there are new things to be experienced in, but few experienced people at first. The trouble is, someone with experience in one or two decades of the same kind of technology in the past are shunned because they can't actually list the new technology now, even though they would probably be up to speed in a week or two (so the employer would rather spend 3 months continuing to look and eventually hire someone on H-1B who has very little experience, but is at the bottom of the range of salary to meet H-1B requirements).
This one looks interesting. A couple of the pages on Toms' site were mangled (bad HTML and picky NS 4) but I got the gist of it. I went to the Shuttle site to see more, but they showed only less. Now to figure out where to get one.
I'm looking for a complete unit, with all hardware, ready to accept my software. But what I am looking for is one that is PC compatible, with space for a CDROM drive or a hard drive, plus 2 NICs, but without being the size of a PC. So far I have not seen anyone accomplish this.
You should not be using the same password for all your sites, even if the authentication mechanism never lets the site server have the actual password. If this one password is exposed by your own accident or something, you've basically given whoever has it access to everything. You might as well hand them your wallet, too.
To track spamming leaks, I also give each place which gets my email address a different one. So there's another piece of information that needs to be different. Not everyone yet has the ability to do this, and not everyone will want to. But a lot of people will unless the spam problem gets solved (unlikely).
Anyway, I see major privacy risks in both Liberty Alliance as well as Passport, particularly in not letting people (easily?) control who gets what information.
Don't forget about the spam that comes along that includes tags which causes your mail reader (in HTML mode) to download those images... possibly without you even knowing. Yes, I have seen web sites that have <img width=1 height=1 src="http://goatse.cx/hello.jpg"> which causes it to be downloaded into a little dot you probably never notice. But if your ISP (or employer as the case may be) is spying on you... oops!
Suppose they decide next to block sites that offer illegal copies of music, movies, and software? I'm sure the RIAA, MPAA, and BSA would applaud such a law. Would you? What would be next? Sites that describe sexual information? Sites that describe which chemicals have explosive reactions? Sites that describe how to fly airplanes?
Routing lookup and filter processing work differently. A routing lookup can be done with a kind of hash, and is often done in hardware for maximum speed. Filter processing is more complicated due to the fact that it has to test more kinds of things, and make varied decisions based on the results. That ends up requiring that the filters be tested in sequence. Unfortunately, the filter matching on addresses are not usually implemented as a hash lookup, and so, each filter access-list entry does one match at a time, in the specified sequence. I've seen routers slow down by having too many access-list entries. This could be designed better in routers and I could describe how, but the sad fact is it hasn't been done anywhere I've seen (most Cisco). But since this kind of blocking isn't the kind needed to keep a DoS attack from going further into the network, it works to simply add the addresses to be blocked to the route table and send their packets to a null interface (e.g. the bit bucket). The web server with the pr0n thus never even gets the SYN packet and no connection is ever established.
If you actually get the list, maybe it will say how. Maybe. Either way, you should definitely post the list so that the rest of us will know what sites to block on our routers, all thanks to the investigative work paid for by the fine taxpayers of Pennsyvania:-)
Apparently from other comments I gather that the PA attorney general provides the list and you block those and are not responsible for having to track changes, except according to a new list they would provide.
I just wonder what form the list will be in. For example, what if they provide it in Microsoft Excel format, and it doesn't load up correctly on {K,Open,Star}Office because of some proprietary thing Microsoft used. I'd sure have some fun with that.
Send me the list and I'll post it. I can't be fired unless I do it myself, and I promise not to fire myself. If you run a network, you know how to figure out my email address.
BTW, this causes collateral damage, the thing that anti-anti-spammers seem to whine about a lot. So either the anti-anti-spammers should come charging in and say this is bad, or else it can be used as a defense for the collateral damage concept many anti-spammers are using.
Sun needs a transition plan to make migration from the low end Linux/x86 based desktops and servers to their Solaris/Sparc based high end workstations and enterprise servers. Otherwise they will not be able to bring as much sales up to the higher tier. There are two ways to do this. One is to run Solaris on x86 hardware as the middle tier. The other is to run Linux on Sparc hardware as the middle tier. One of these approaches leaves Sun subject to the whims of another CPU maker, which has it's own plans for 64-bitdomination. The other leaves Sun subject to the whims of a huge open source software community and a few choices in Linux distributions (such as Debian, Mandrake, and SuSE) as well as FreeBSD, NetBSD, and OpenBSD. Which way do you think would be better for Sun?
With enough people lined up across a big country, with very accurate clocks to know when to do it, they can raise and lower their arms to create a wave that goes faster than the speed of light. But this doesn't mean anything. Suppose you have 2 people at each end of town (assuming a big enough town... 1 km is plenty) doing it, and they do it a microsecond apart. Does that mean something traveled faster than the speed of light? No. They just timed things well. If you stand near the person doing it late, you see them first and the other next. Does that mean time went backwards? Of course not. It just means your "order of influence" is reversed because of your position.
Set up a series of ham radio beacon transmitters spaced at intervals greater than 187 miles (300 km), syncronized in time to better than a microsecond, scheduled to transmit a half millisecond RF burst at one millisecond intervals going from east to west. Give each a slightly different frequency. The listener at the west end will hear the pulse from the closer one first then progressively from the more distant beacon. Time still didn't go backwards.
Now if you were standing at the end where the laser (or bullet, in another poster's machine gun analogy) impacts are coming to, what would it look like to you (assuming it stops just short of hitting you)? The answer is, you'd see the closer impacts first, and the more distant impacts later. It would appear that they are going away from you. So from this perspective, time would appear to be going backwards.
The thing is, we might actually see such things happen out in space. Stars that are emitting energy in a specific direction, other than their poles, and are rotating, can illuminate dust clouds at some distance off to the side. On the side where the rotation is coming towards us, and at a distance sufficient to make the effect traverse faster than light, we'll actually see (if we can see that level of resolution) the effect go backwards. Combining the effect with an accurate rotation rate measurement, a very accurate distance from the star to the dust cloud can be measured. Then from there you can work back to an accurate mesure of the distance. In reality the distances will be rather small for quickly rotating stars, so it can't be observed directly. But surely it's effects can be predicted from other determinations of that distance and rotation rate, and then used to confirm those measurements.
Anyone have a mirror of the NewScientist web site? Their web programmer is clueless (and has been told about this a few times) and developing stuff that is incompatible with some proxy servers.
Ooops! Sorry, I dropped it. But don't worry, I didn't lose any parts. I glued it all back together. Ignore those minor cosmetic blemishes. And it might skip tracks occaisionally.
You don't even need to cut the wires. You can just put a coil around the earpieces or the wires leading to the earpieces and pick up the content inductively. Most journalists won't know that, but it only takes one leak:-)
I built a large scale shopping cart style web site back in 1996 that worked on browsers as far back as Netscape 2. It produced pure HTML. I was the programmer and I set it up so the web designer guys could control the entire look and feel of the pages. Now they didn't have control over the database logic and how things were looked up, but they did have control over the basics, like page layout, fonts, colors, images, buttons, menus, and such. And it was done using a prototype page. They simply created an HTML page using any HTML editor they liked, which had some specific elements in it, like the words "menu here" where they wanted the menu to go. My shopping cart code simply took the prototype template and extracted the elements and used them to output HTML where it needed to. They changed the entire site design 3 times while I worked there and the shopping cart designed changed right along with it. And all this was without any CSS. Ironically, if they had added CSS to the prototype, it would have ended up on the shopping cart site pages. The fun part was coming in each day and checking the site myself to see if they had made a change. I never had to do anything to support it. And it simply used HTML. So HTML can be done on a large dynamic site and without the programmer having to deal with all the changes.
Well adding CSS to HTML is what I meant to say is better. Your argument is based on that. And I even agree to that. I don't say that is wrong because I believe it is right. CSS is a good thing to have. My point is still that we are both seeking the same end goal but your path to get there is the one that involves jumping over the cliff, while mine is to climb down the side.
If I want to change my HTML, I'll change it. Of course without CSS, I have to change the HTML instead. No big deal. It's not carved in stone. Maybe you just don't get HTML. Of course HTML wasn't designed for this, and CSS is more flexible for certain things that HTML is poor at. But again, I really can change the HTML and it's not that much trouble.
If I want the text to be a different size, I can in fact change it with <font size=+1> or the like. Of course this disobeys the principle that look and feel (which would include size of text based on properties like what it is for, and what the user's theme is) should be decided at the end point. But CSS breaks that rule, too, to the extent that the user cannot substitute their own stylesheet to override the web server provided one (which I know of no browser that can do this properly). And of course, I do use things like <h1> where appropriate.
You're arguments are based on the idea that CSS is better than HTML because it is appropriate for the task. And I don't deny that. But your arguments go further and say that HTML cannot even do things that I actually can make it do. So I know that aspect of your argument is without basis.
The real problem is that not everyone has upgraded their browsers to the point where CSS can be relied on. The choices at that point are:
Use CSS alone, which results in totally screwed up layouts for lots of people.
Use CSS and HTML in conjunction with each other, which lots of developers object to because it is redundant. But at least it works optimally for everyone.
Use HTML alone for now, without any CSS, which still works for everyone, but is less bandwidth and work than the combination approach.
I happen to choose the 3rd approach. I'll probably never do the 2nd. I'll switch to the 1st when a sufficient percentage of people use a browser that correctly supports CSS.
But what about this upgrading? I know why it is that I don't. But I do see browsers eventually making it to the point where the issues I have will no longer be of sufficient concern to preclude the switch. It is a balancing act. No browser is perfect. I'm trading off one set of issues for another set when I change browsers. And right now for me, Netscape version 4 has fewer issues than any of the others. Konquerer 3 is poised to edge it out, perhaps in the next minor version, or when I have the time to iron out some of the startup configuration issues myself (I have a couple ideas of a hack to get around enough issues to make the switch a gain).
But even if I upgrade, whether I switch to CSS in my web pages or not depends on how many others upgrade. My reason for not upgrading is not necessarily (and almost certainly not) the reason others don't upgrade. I'll leave their reasons up to them. But I do know that the new browsers have a lot of failings, especially in the area of using so much more CPU time to render, and using so much more RAM (not just mapping it, either... actually touching it, so the run-time foot print is substantially larger).
As long as developers who want to use CSS and other new standards focus on trying to get people to upgrade without focusing on trying to get good browsers that people can upgrade to (and there is no one size fits all), then you just simply will not see the results you want. I've explained this to web developers many times, and they still just don't seem to get it. Quite many people don't do upgrades just because something is new. They upgrade when it means they will be gaining more than they lose. And there is always a loss in upgrading even to the perfect browser (which does not exist), that being their (or other staff, in a business) time. But really, there are other losses in new browsers to overcome. What else can I say to explain to you that which you still don't understand?
Maybe you can get everyone to upgrade by finding a dangerously exploitable security flaw in all the older browsers:-)
Slashdot Mirror
There are currently thousands of highly qualified people available now who will work for half or even a third of the salary as average. This is why the government conspired with wall street to bust the big bubble, because no one would work for the government anymore (no stock options). And unlike H-1Bs, who have to be paid what the average person makes, you can legally pay Americans way under average. So now there are plenty of admins available ... and programmers, too. Just post the openings here and watch the geeks resumes come flooding in.
My message would be the source code to DeCSS, compressed, and alpha-encoded.
There are lots of jobs like that in many companies. Instead of having one person do it for all the branches of various different companies in the same city (e.g. via a contracting firm that arranges it this way), they end up shuttling people all around. While one person flies from Dallas to Atlanta to install XP, another flies from Atlanta to Dallas to install XP. Well, I guess it helps the airlines; they need it.
There are good apples and bad apples in every bunch. I've met both good and bad H-1B people. I remember going on a sales visit (four of us, including our director of sales) to a major telco equipment manufacturer to demonstrate a software product. We were showing this to a couple of managers who were assisted by their lead system administrator (who ran a network of servers and workstations for an engineering group of 300 people). I believe there was one other sysadmin working there, but he was not in this meeting. While the sysadmin managed to install the package just fine, he wasn't able to configure the network on the machine (Solaris) to successfully talk to the clients. So I had to tell him what commands to type in. There were a few other cases of him not knowing what he was doing. And the two managers didn't seem to be surprised at this at all. Toward the end of the meeting I found out he was an H-1B worker from Egypt.
For every example someone can come up with for a good person, there will be another for a bad person. For every example someone can come up with for a bad person, there will be another for a good person. But what is a fact is that the US government simply does not do a thorough review of the H-1B applications. They don't even have staff qualified to understand the various sub fields that technology employers need. They just make sure the paperwork is right and the named person isn't a known terrorist and stamp "approved". The whole process for H-1B costs a fraction of what the payout to a technology talent recruiter would make placing just one really good techie.
During the boom, sure, it was hard to find people, but that wasn't from lack of people. It was from a non-streamlined recruiting process that in itself was very expensive, and doing a poor job of communicating and matching people (I was frequently called up for programming jobs, and that was not what my resume said I was looking for ... either they can't read or they were extremely desperate). Even the job board web sites suck today. Put in "linux" as a keyword and you match lots of Windows jobs that say "experience with linux a plus".
YOUR company may not be abusing the system, but many companies are. One recent example exposed was Bank of America. Not only did they replace lots of IT workers with H-1B people, they even required the replaced domestic workers to stay a few weeks and train their H-1B replacements in order to qualify for the severance package. This is why I refuse to do business with BoA.
What makes you think that H-1B abuses get caught? The government isn't reviewing them. The companies doing the abuse certainly aren't telling. Yet people are being replaced by H-1B workers in both the boom and the bust.
It's the largest corporations that have the H-1B process streamlined where it's no longer a hassle for them.
And going back to school is not the answer. What would you do, get a 2nd CS degree to replace your first? If you go back to school I recommend getting a degree in Salesmanship ... there is a current shortage of good sales people. What high tech businesses want in their employees is experience. Part of the problem where shortages exist is that there is less of a pathway to achieve experience than there has been before. Even during the boom, less experienced and inexperienced people could not find jobs (I know some personally who had this trouble). Another part of the problem is that as technology changes, there are new things to be experienced in, but few experienced people at first. The trouble is, someone with experience in one or two decades of the same kind of technology in the past are shunned because they can't actually list the new technology now, even though they would probably be up to speed in a week or two (so the employer would rather spend 3 months continuing to look and eventually hire someone on H-1B who has very little experience, but is at the bottom of the range of salary to meet H-1B requirements).
No price. No "add to shopping cart" button.
This one looks interesting. A couple of the pages on Toms' site were mangled (bad HTML and picky NS 4) but I got the gist of it. I went to the Shuttle site to see more, but they showed only less. Now to figure out where to get one.
I'm looking for a complete unit, with all hardware, ready to accept my software. But what I am looking for is one that is PC compatible, with space for a CDROM drive or a hard drive, plus 2 NICs, but without being the size of a PC. So far I have not seen anyone accomplish this.
And where can I buy this 3 NIC version today?
You should not be using the same password for all your sites, even if the authentication mechanism never lets the site server have the actual password. If this one password is exposed by your own accident or something, you've basically given whoever has it access to everything. You might as well hand them your wallet, too.
To track spamming leaks, I also give each place which gets my email address a different one. So there's another piece of information that needs to be different. Not everyone yet has the ability to do this, and not everyone will want to. But a lot of people will unless the spam problem gets solved (unlikely).
Anyway, I see major privacy risks in both Liberty Alliance as well as Passport, particularly in not letting people (easily?) control who gets what information.
Don't forget about the spam that comes along that includes tags which causes your mail reader (in HTML mode) to download those images ... possibly without you even knowing. Yes, I have seen web sites that have <img width=1 height=1 src="http://goatse.cx/hello.jpg"> which causes it to be downloaded into a little dot you probably never notice. But if your ISP (or employer as the case may be) is spying on you ... oops!
Suppose they decide next to block sites that offer illegal copies of music, movies, and software? I'm sure the RIAA, MPAA, and BSA would applaud such a law. Would you? What would be next? Sites that describe sexual information? Sites that describe which chemicals have explosive reactions? Sites that describe how to fly airplanes?
Routing lookup and filter processing work differently. A routing lookup can be done with a kind of hash, and is often done in hardware for maximum speed. Filter processing is more complicated due to the fact that it has to test more kinds of things, and make varied decisions based on the results. That ends up requiring that the filters be tested in sequence. Unfortunately, the filter matching on addresses are not usually implemented as a hash lookup, and so, each filter access-list entry does one match at a time, in the specified sequence. I've seen routers slow down by having too many access-list entries. This could be designed better in routers and I could describe how, but the sad fact is it hasn't been done anywhere I've seen (most Cisco). But since this kind of blocking isn't the kind needed to keep a DoS attack from going further into the network, it works to simply add the addresses to be blocked to the route table and send their packets to a null interface (e.g. the bit bucket). The web server with the pr0n thus never even gets the SYN packet and no connection is ever established.
And yes, there are ways around it. 99% of the masses will never even think to try to go around it, which is probably sufficient to satisfy the Pennsylvania Attorney General's Office Criminal Law Division Child Sexual Exploitation Unit.
If you actually get the list, maybe it will say how. Maybe. Either way, you should definitely post the list so that the rest of us will know what sites to block on our routers, all thanks to the investigative work paid for by the fine taxpayers of Pennsyvania :-)
Apparently from other comments I gather that the PA attorney general provides the list and you block those and are not responsible for having to track changes, except according to a new list they would provide.
I just wonder what form the list will be in. For example, what if they provide it in Microsoft Excel format, and it doesn't load up correctly on {K,Open,Star}Office because of some proprietary thing Microsoft used. I'd sure have some fun with that.
I'm sure they know which routers are in PA. They can do the null route thing just on those.
Send me the list and I'll post it. I can't be fired unless I do it myself, and I promise not to fire myself. If you run a network, you know how to figure out my email address.
BTW, this causes collateral damage, the thing that anti-anti-spammers seem to whine about a lot. So either the anti-anti-spammers should come charging in and say this is bad, or else it can be used as a defense for the collateral damage concept many anti-spammers are using.
And spammers are restricted to just the 3rd section.
Don't worry, the net will be safe for the next 10-15 minutes while all the hackers go get their laughs.
Sun needs a transition plan to make migration from the low end Linux/x86 based desktops and servers to their Solaris/Sparc based high end workstations and enterprise servers. Otherwise they will not be able to bring as much sales up to the higher tier. There are two ways to do this. One is to run Solaris on x86 hardware as the middle tier. The other is to run Linux on Sparc hardware as the middle tier. One of these approaches leaves Sun subject to the whims of another CPU maker, which has it's own plans for 64-bit domination. The other leaves Sun subject to the whims of a huge open source software community and a few choices in Linux distributions (such as Debian, Mandrake, and SuSE) as well as FreeBSD, NetBSD, and OpenBSD. Which way do you think would be better for Sun?
With enough people lined up across a big country, with very accurate clocks to know when to do it, they can raise and lower their arms to create a wave that goes faster than the speed of light. But this doesn't mean anything. Suppose you have 2 people at each end of town (assuming a big enough town ... 1 km is plenty) doing it, and they do it a microsecond apart. Does that mean something traveled faster than the speed of light? No. They just timed things well. If you stand near the person doing it late, you see them first and the other next. Does that mean time went backwards? Of course not. It just means your "order of influence" is reversed because of your position.
Set up a series of ham radio beacon transmitters spaced at intervals greater than 187 miles (300 km), syncronized in time to better than a microsecond, scheduled to transmit a half millisecond RF burst at one millisecond intervals going from east to west. Give each a slightly different frequency. The listener at the west end will hear the pulse from the closer one first then progressively from the more distant beacon. Time still didn't go backwards.
Now if you were standing at the end where the laser (or bullet, in another poster's machine gun analogy) impacts are coming to, what would it look like to you (assuming it stops just short of hitting you)? The answer is, you'd see the closer impacts first, and the more distant impacts later. It would appear that they are going away from you. So from this perspective, time would appear to be going backwards.
The thing is, we might actually see such things happen out in space. Stars that are emitting energy in a specific direction, other than their poles, and are rotating, can illuminate dust clouds at some distance off to the side. On the side where the rotation is coming towards us, and at a distance sufficient to make the effect traverse faster than light, we'll actually see (if we can see that level of resolution) the effect go backwards. Combining the effect with an accurate rotation rate measurement, a very accurate distance from the star to the dust cloud can be measured. Then from there you can work back to an accurate mesure of the distance. In reality the distances will be rather small for quickly rotating stars, so it can't be observed directly. But surely it's effects can be predicted from other determinations of that distance and rotation rate, and then used to confirm those measurements.
Anyone have a mirror of the NewScientist web site? Their web programmer is clueless (and has been told about this a few times) and developing stuff that is incompatible with some proxy servers.
Ooops! Sorry, I dropped it. But don't worry, I didn't lose any parts. I glued it all back together. Ignore those minor cosmetic blemishes. And it might skip tracks occaisionally.
You don't even need to cut the wires. You can just put a coil around the earpieces or the wires leading to the earpieces and pick up the content inductively. Most journalists won't know that, but it only takes one leak :-)
I built a large scale shopping cart style web site back in 1996 that worked on browsers as far back as Netscape 2. It produced pure HTML. I was the programmer and I set it up so the web designer guys could control the entire look and feel of the pages. Now they didn't have control over the database logic and how things were looked up, but they did have control over the basics, like page layout, fonts, colors, images, buttons, menus, and such. And it was done using a prototype page. They simply created an HTML page using any HTML editor they liked, which had some specific elements in it, like the words "menu here" where they wanted the menu to go. My shopping cart code simply took the prototype template and extracted the elements and used them to output HTML where it needed to. They changed the entire site design 3 times while I worked there and the shopping cart designed changed right along with it. And all this was without any CSS. Ironically, if they had added CSS to the prototype, it would have ended up on the shopping cart site pages. The fun part was coming in each day and checking the site myself to see if they had made a change. I never had to do anything to support it. And it simply used HTML. So HTML can be done on a large dynamic site and without the programmer having to deal with all the changes.
Well adding CSS to HTML is what I meant to say is better. Your argument is based on that. And I even agree to that. I don't say that is wrong because I believe it is right. CSS is a good thing to have. My point is still that we are both seeking the same end goal but your path to get there is the one that involves jumping over the cliff, while mine is to climb down the side.
If I want to change my HTML, I'll change it. Of course without CSS, I have to change the HTML instead. No big deal. It's not carved in stone. Maybe you just don't get HTML. Of course HTML wasn't designed for this, and CSS is more flexible for certain things that HTML is poor at. But again, I really can change the HTML and it's not that much trouble.
If I want the text to be a different size, I can in fact change it with <font size=+1> or the like. Of course this disobeys the principle that look and feel (which would include size of text based on properties like what it is for, and what the user's theme is) should be decided at the end point. But CSS breaks that rule, too, to the extent that the user cannot substitute their own stylesheet to override the web server provided one (which I know of no browser that can do this properly). And of course, I do use things like <h1> where appropriate.
You're arguments are based on the idea that CSS is better than HTML because it is appropriate for the task. And I don't deny that. But your arguments go further and say that HTML cannot even do things that I actually can make it do. So I know that aspect of your argument is without basis.
The real problem is that not everyone has upgraded their browsers to the point where CSS can be relied on. The choices at that point are:
- Use CSS alone, which results in totally screwed up layouts for lots of people.
- Use CSS and HTML in conjunction with each other, which lots of developers object to because it is redundant. But at least it works optimally for everyone.
- Use HTML alone for now, without any CSS, which still works for everyone, but is less bandwidth and work than the combination approach.
I happen to choose the 3rd approach. I'll probably never do the 2nd. I'll switch to the 1st when a sufficient percentage of people use a browser that correctly supports CSS.But what about this upgrading? I know why it is that I don't. But I do see browsers eventually making it to the point where the issues I have will no longer be of sufficient concern to preclude the switch. It is a balancing act. No browser is perfect. I'm trading off one set of issues for another set when I change browsers. And right now for me, Netscape version 4 has fewer issues than any of the others. Konquerer 3 is poised to edge it out, perhaps in the next minor version, or when I have the time to iron out some of the startup configuration issues myself (I have a couple ideas of a hack to get around enough issues to make the switch a gain).
But even if I upgrade, whether I switch to CSS in my web pages or not depends on how many others upgrade. My reason for not upgrading is not necessarily (and almost certainly not) the reason others don't upgrade. I'll leave their reasons up to them. But I do know that the new browsers have a lot of failings, especially in the area of using so much more CPU time to render, and using so much more RAM (not just mapping it, either ... actually touching it, so the run-time foot print is substantially larger).
As long as developers who want to use CSS and other new standards focus on trying to get people to upgrade without focusing on trying to get good browsers that people can upgrade to (and there is no one size fits all), then you just simply will not see the results you want. I've explained this to web developers many times, and they still just don't seem to get it. Quite many people don't do upgrades just because something is new. They upgrade when it means they will be gaining more than they lose. And there is always a loss in upgrading even to the perfect browser (which does not exist), that being their (or other staff, in a business) time. But really, there are other losses in new browsers to overcome. What else can I say to explain to you that which you still don't understand?
Maybe you can get everyone to upgrade by finding a dangerously exploitable security flaw in all the older browsers :-)