No, I'm not presupposing that people are out buying Windows specifically to view some SL only sites. I'm conjecturing that if Microsoft wants to abuse their monopoly, they might make SL work in Windows only, and then people would buy Windows specifically to view some SL only sites.
I agree that switching a site from IE-only to working in other browsers is expensive. Any time you need to switch from one technology to another, it's very expensive and risky. That's why businesses continue buying Windows machines instead of switching to Linux. Switching to Linux would mean having to run a browser other than IE, an office suite other than Office, an email client other than Outlook, etc. Making each of these switches has interoperability problems, because of the proprietary extensions in IE, and the proprietary nature of Office and the Outlook/Exchange protocol. It's called vendor lock-in, and it makes MS tons of bucks. If MS products used and adhered closely to standards, they wouldn't have the lock-in that they do now. Silverlight is yet another proprietary technology that can be used to induce yet more lock-in.
Bundling is certainly one way to abuse a monopoly. Another way is limiting interoperability. It's especially powerful when the two are used together. The ability to share calendars in Outlook requires Exchange on the server side. By bundling Outlook into Office and trial versions of Office with Windows, users get exposed to Outlook. Then they find out they need to run Exchange to share their calendars, which requires a Windows Server. By default, Exchange uses MAPI to communicate with email clients, so all users who connect to the Exchange server find they need to use Outlook, which requires Windows on the desktop.
Similarly, Microsoft bundling IE with Windows caused the usage of IE to go so high that some developers wrote sites that work only in IE. To access those sites, users now find they need to run Windows to run IE so they can access those sites.
You're woefully naive if you think Microsoft is in the business of creating products that compete on a level playing field with products from other companies. They are well skilled at using bundling and limiting interoperability to lock users into other Microsoft products. I note that Silverlight is not compatible with other multimedia players and will be bundled with Windows. Hmmm...
You are correct that if Microsoft stopped developing Silverlight for Firefox, developers would abandon Silverlight. However, developers wouldn't be able to switch overnight, and in the meantime Microsoft would be making money by selling Windows to users can use IE and Silverlight to view the sites that still use Silverlight. Similarly, many sites (especially intranet sites) work only in IE. Microsoft is still making money by selling Windows to those pour souls who still have to access IE-only sites, even if no developers are developing new sites to work only in IE.
I never said Microsoft should have come out with Silverlight, or that competition is bad. That would be silly. I was just making the point that Silverlight is incompatible with Flash at the source code level. It's not like a different compiler or a different player for the same language.
Well, you can argue Flash has a monopoly now for more interactive sites. What if they stop making a plugin for FF?
Then web developers will stop using Flash, as it doesn't work for more than 20% their visitors. Then Adobe loses their monopoly and their power.
On the other hand, if Microsoft stops developing Silverlight for Firefox, will people stop buying Windows? I just don't understand how people still don't get it. Microsoft has a monopoly in the desktop OS market. They abuse that monopoly in other areas, for example, attempting to lock people in to office suites (Office), browsers (IE), languages (C#), servers (Windows server and Exchange).
So if this was MS trying to lock people in, it would be Windows and IE only, instead it has potential to be far more crossplatform than Flash.
If I were trying to lock people in, I would develop the technology for all popular platforms at first. After it became very popular, I would slowly drop support for platforms other than my own, first Linux, then Mac, then non-IE browsers.
Well, AMD isn't stiving to make the CPU world a better place, they are trying to beat Intel. AMD would love to get all of Intel's marketshare, I'm sure, and Intel feels the same way. What exactly is wrong with that?
Nothing. But because Microsoft is a monopolist that has in the past abused their monopoly power, I would be wary of new technologies the produce. What if they stop making the player for operating systems other than Windows when Silverlight becomes popular. What if they stop making a player for browsers other than IE? Remember, embrace, extent, extinguish.
A different compiler for the same language is one thing. Silverlight is like a completely new language. This is like Microsoft introducing C# when we already have C++ and Java.
More to the point, why should anyone who has a website develop content for Silverlight instead of Flash? Flash is already well deployed, and Adobe has an interest in maintaining Flash players for multiple operating systems. Silverlight is not only barely deployed, but would seem that Microsoft would have an interest in developing players that run only on Windows, and perhaps only on Internet Explorer, once Silverlight becomes popular.
Keep in mind the original question. The submitter had problems finding a developer for their language. Are there really as many Python, Ruby, or TCL programmers out there are there are Perl, PHP, or ASP programmers? That's why I tend to stick with more commonly used languages. I haven't bothered to learn Python or Ruby, as I prefer to use Perl in case I need to hand the code off to someone else. I write server-side web code in PHP, even though I'd rather stick with Perl, which I'm more familiar with.
NONMEM, used by most (if not all) scientists doing pharmacokinetics/pharmacodynamics, is written in FORTRAN 77. I'm sure there are many more examples of important scientific and engineering code still maintained in FORTRAN. When gfortran becomes about as stable as g77 was, perhaps most will really move on to Fortran 95.
It's easy to generate some sorts of tests using a random-number generator, actually...
Yes, fuzz testing can be very useful at finding obscure bugs that no web developer would knowingly trigger. They're useful for finding edge cases where extreme bugs such as crashes and security problems can occur. That's not what Acid3 is testing, however.
Acid3 is pointing out deficiencies in current browsers, as the OP conjectured. But it is not pointing out those deficiencies for unfair or useless browser bashing. It is pointing the deficiencies out that make it harder for web developers to develop Web 2.0 applications.
Firefox and Opera render accurately, too. Firefox used to be a JavaScript slowpoke, but with the latest performance improvements, Firefox 3 is fast at JavaScript, too. I'm not trying to advocate one browser over another, just point out that there are at least three good, fast, popular browsers out there. It doesn't seem like one is generally better or worse than the others overall.
Both Firefox and Safari developers have obviously been fixing bugs that Acid3 demonstrates for the past few months. The Firefox tracking bug for Acid3 problems is over two months old, for example. If you want to see how well the browsers did before the test existed, use one from April 2007, when work on Acid3 started.
I don't know if Safari is the most standards compliant either. Out of the latest stable release versions of Firefox, Opera, and Safari, Safari does the worst on Acid3. On the other hand, out of the latest development versions of those browsers, Safari does best. It looks to me like Safari passed the Acid2 test quickly, and will pass the Acid3 test quickly, not because Safari is standards king, but because Safari developers work quickly on making Safari pass Acid tests. And remember, how well browsers do on Acid tests is not any kind of measure of how good the browser is.
Firefox, Safari, and Opera all have very good standards support, and some are better than others at one or another area of web standards, but no one browser stands out as generally better at standards than the others.
In a way, yes it is a waste. But when 90% of the pages on the web don't validate, it is a necessity. Your fellow web developers don't seem to be as diligent as you are, and they do care what browsers do with their buggy code, as do web users. It seems as if web developers such as yourself are the only people who don't care, and you're a tiny minority in the big equation.
Yeah, for every item tested by Acid3, I am sure there exists a web developer who would have benefitted from the issue being fixed. If no web developers wanted a feature, it would not have been written into a web standard. If no web developers ran into a bug, they would not have known to test for it.
The bottom line is that the OP is correct. The scores are not a meaningful indicator of "goodness" of a browser. They simply indicate how many tests each browser passed. The purpose of the test is not to make browsers look good or bad, but to ensure web browsers support the features requested by web developers for Web 2.0. As the description for the Acid3 test states, Acid3 is primarily testing specifications for "Web 2.0" dynamic Web applications.
If you look closely at the results people are testing, Webkit wasn't doing so well when the Acid3 test was first becoming available. If you look at the most recent stable versions, Safari 3.0.4 scores only 39/100, but Firefox 2.0.0.12 scores 49/100. In that latest development version, Safari scores 90/100, but Firefox scores only 67/100. The Webkit developers are obviously deliberately fixing the bugs that Acid3 demonstrates. That great and all, but the high score doesn't mean it's an inherently superior rendering engine. It just means the Safari developers work quickly and hard to get a good score on the Acid tests, just as they did with Acid2. The scores browsers get on Acid tests really have no meaning except for how well the browser does on the Acid test, which is often just an indication of how much developers have worked on getting the browser to pass the Acid test.
When was Opera ever standards king? Before Opera 7 I couldn't use Opera because it couldn't reflow web pages correctly to handle DHTML. The Mozilla and IE of the day could handle reflow just fine. I couldn't use Opera 8 because it didn't support XSLT, and one website that I frequented used XSLT. The IE and Firefox of the day could handle XSLT just fine. Opera's got good standards support, generally about as good as Firefox or Safari, but it seems like it's often playing catch-up in at least one area. But standards king? I think not.
Of course developers should care how browsers deal with incorrect stuff. Most web pages today have incorrect stuff. If all browsers handle the incorrect stuff differently, they'll all show the pages in different ways. If we standardize how to handle incorrect stuff, web pages can look the same in every browser even if they are not correctly made.
Re:How do the acid-test creators test the acid tes
on
Acid3 Test Released
·
· Score: 1
I've never heard of database theory or computation theory as first year courses in CS. As far as I know, they mostly teach introduction to programming, algorithms, and data structures in the first year. That's what Computer Science AP classes teach, also.
You're right... many of the Acid3 tests use assert and assertEquals to test that results are as expected, and these methods throw an exception when they aren't.
You're exactly correct that the Acid tests test specific browser flaws. They are testing exactly the flaws that plague web developers. That way, when all popular browsers pass the Acid tests, web developers don't need to work around the flaws in each different browser. We all benefit by getting web sites with fancy new features that work in all browsers. The scores are not meaningful, but are a way to motivate the developers of web browsers to fix their flaws so they're not embarrassed by a low or non-passing score.
Slashdot Mirror
No, I'm not presupposing that people are out buying Windows specifically to view some SL only sites. I'm conjecturing that if Microsoft wants to abuse their monopoly, they might make SL work in Windows only, and then people would buy Windows specifically to view some SL only sites.
I agree that switching a site from IE-only to working in other browsers is expensive. Any time you need to switch from one technology to another, it's very expensive and risky. That's why businesses continue buying Windows machines instead of switching to Linux. Switching to Linux would mean having to run a browser other than IE, an office suite other than Office, an email client other than Outlook, etc. Making each of these switches has interoperability problems, because of the proprietary extensions in IE, and the proprietary nature of Office and the Outlook/Exchange protocol. It's called vendor lock-in, and it makes MS tons of bucks. If MS products used and adhered closely to standards, they wouldn't have the lock-in that they do now. Silverlight is yet another proprietary technology that can be used to induce yet more lock-in.
Bundling is certainly one way to abuse a monopoly. Another way is limiting interoperability. It's especially powerful when the two are used together. The ability to share calendars in Outlook requires Exchange on the server side. By bundling Outlook into Office and trial versions of Office with Windows, users get exposed to Outlook. Then they find out they need to run Exchange to share their calendars, which requires a Windows Server. By default, Exchange uses MAPI to communicate with email clients, so all users who connect to the Exchange server find they need to use Outlook, which requires Windows on the desktop.
Similarly, Microsoft bundling IE with Windows caused the usage of IE to go so high that some developers wrote sites that work only in IE. To access those sites, users now find they need to run Windows to run IE so they can access those sites.
You're woefully naive if you think Microsoft is in the business of creating products that compete on a level playing field with products from other companies. They are well skilled at using bundling and limiting interoperability to lock users into other Microsoft products. I note that Silverlight is not compatible with other multimedia players and will be bundled with Windows. Hmmm...
You are correct that if Microsoft stopped developing Silverlight for Firefox, developers would abandon Silverlight. However, developers wouldn't be able to switch overnight, and in the meantime Microsoft would be making money by selling Windows to users can use IE and Silverlight to view the sites that still use Silverlight. Similarly, many sites (especially intranet sites) work only in IE. Microsoft is still making money by selling Windows to those pour souls who still have to access IE-only sites, even if no developers are developing new sites to work only in IE.
And what happens when they come out with an incompatible new version of Silverlight and don't release the documentation and don't help with the development of the FOSS version? From what I've seen, developers are still leery of Microsoft's recent "Interoperability Initiative".
I never said Microsoft should have come out with Silverlight, or that competition is bad. That would be silly. I was just making the point that Silverlight is incompatible with Flash at the source code level. It's not like a different compiler or a different player for the same language.
Then web developers will stop using Flash, as it doesn't work for more than 20% their visitors. Then Adobe loses their monopoly and their power.
On the other hand, if Microsoft stops developing Silverlight for Firefox, will people stop buying Windows? I just don't understand how people still don't get it. Microsoft has a monopoly in the desktop OS market. They abuse that monopoly in other areas, for example, attempting to lock people in to office suites (Office), browsers (IE), languages (C#), servers (Windows server and Exchange).
A different compiler for the same language is one thing. Silverlight is like a completely new language. This is like Microsoft introducing C# when we already have C++ and Java.
More to the point, why should anyone who has a website develop content for Silverlight instead of Flash? Flash is already well deployed, and Adobe has an interest in maintaining Flash players for multiple operating systems. Silverlight is not only barely deployed, but would seem that Microsoft would have an interest in developing players that run only on Windows, and perhaps only on Internet Explorer, once Silverlight becomes popular.
Keep in mind the original question. The submitter had problems finding a developer for their language. Are there really as many Python, Ruby, or TCL programmers out there are there are Perl, PHP, or ASP programmers? That's why I tend to stick with more commonly used languages. I haven't bothered to learn Python or Ruby, as I prefer to use Perl in case I need to hand the code off to someone else. I write server-side web code in PHP, even though I'd rather stick with Perl, which I'm more familiar with.
NONMEM, used by most (if not all) scientists doing pharmacokinetics/pharmacodynamics, is written in FORTRAN 77. I'm sure there are many more examples of important scientific and engineering code still maintained in FORTRAN. When gfortran becomes about as stable as g77 was, perhaps most will really move on to Fortran 95.
Yes, fuzz testing can be very useful at finding obscure bugs that no web developer would knowingly trigger. They're useful for finding edge cases where extreme bugs such as crashes and security problems can occur. That's not what Acid3 is testing, however.
Acid3 is pointing out deficiencies in current browsers, as the OP conjectured. But it is not pointing out those deficiencies for unfair or useless browser bashing. It is pointing the deficiencies out that make it harder for web developers to develop Web 2.0 applications.
Firefox and Opera render accurately, too. Firefox used to be a JavaScript slowpoke, but with the latest performance improvements, Firefox 3 is fast at JavaScript, too. I'm not trying to advocate one browser over another, just point out that there are at least three good, fast, popular browsers out there. It doesn't seem like one is generally better or worse than the others overall.
Both Firefox and Safari developers have obviously been fixing bugs that Acid3 demonstrates for the past few months. The Firefox tracking bug for Acid3 problems is over two months old, for example. If you want to see how well the browsers did before the test existed, use one from April 2007, when work on Acid3 started.
I don't know if Safari is the most standards compliant either. Out of the latest stable release versions of Firefox, Opera, and Safari, Safari does the worst on Acid3. On the other hand, out of the latest development versions of those browsers, Safari does best. It looks to me like Safari passed the Acid2 test quickly, and will pass the Acid3 test quickly, not because Safari is standards king, but because Safari developers work quickly on making Safari pass Acid tests. And remember, how well browsers do on Acid tests is not any kind of measure of how good the browser is.
Firefox, Safari, and Opera all have very good standards support, and some are better than others at one or another area of web standards, but no one browser stands out as generally better at standards than the others.
In a way, yes it is a waste. But when 90% of the pages on the web don't validate, it is a necessity. Your fellow web developers don't seem to be as diligent as you are, and they do care what browsers do with their buggy code, as do web users. It seems as if web developers such as yourself are the only people who don't care, and you're a tiny minority in the big equation.
Yeah, for every item tested by Acid3, I am sure there exists a web developer who would have benefitted from the issue being fixed. If no web developers wanted a feature, it would not have been written into a web standard. If no web developers ran into a bug, they would not have known to test for it.
The bottom line is that the OP is correct. The scores are not a meaningful indicator of "goodness" of a browser. They simply indicate how many tests each browser passed. The purpose of the test is not to make browsers look good or bad, but to ensure web browsers support the features requested by web developers for Web 2.0. As the description for the Acid3 test states, Acid3 is primarily testing specifications for "Web 2.0" dynamic Web applications.
If you look closely at the results people are testing, Webkit wasn't doing so well when the Acid3 test was first becoming available. If you look at the most recent stable versions, Safari 3.0.4 scores only 39/100, but Firefox 2.0.0.12 scores 49/100. In that latest development version, Safari scores 90/100, but Firefox scores only 67/100. The Webkit developers are obviously deliberately fixing the bugs that Acid3 demonstrates. That great and all, but the high score doesn't mean it's an inherently superior rendering engine. It just means the Safari developers work quickly and hard to get a good score on the Acid tests, just as they did with Acid2. The scores browsers get on Acid tests really have no meaning except for how well the browser does on the Acid test, which is often just an indication of how much developers have worked on getting the browser to pass the Acid test.
When was Opera ever standards king? Before Opera 7 I couldn't use Opera because it couldn't reflow web pages correctly to handle DHTML. The Mozilla and IE of the day could handle reflow just fine. I couldn't use Opera 8 because it didn't support XSLT, and one website that I frequented used XSLT. The IE and Firefox of the day could handle XSLT just fine. Opera's got good standards support, generally about as good as Firefox or Safari, but it seems like it's often playing catch-up in at least one area. But standards king? I think not.
Of course developers should care how browsers deal with incorrect stuff. Most web pages today have incorrect stuff. If all browsers handle the incorrect stuff differently, they'll all show the pages in different ways. If we standardize how to handle incorrect stuff, web pages can look the same in every browser even if they are not correctly made.
Ian Hickson actually did ask for additional tests that fail in Webkit or Firefox, so they did partly write the test against at least two particular browsers.
I've never heard of database theory or computation theory as first year courses in CS. As far as I know, they mostly teach introduction to programming, algorithms, and data structures in the first year. That's what Computer Science AP classes teach, also.
You're exactly correct that the Acid tests test specific browser flaws. They are testing exactly the flaws that plague web developers. That way, when all popular browsers pass the Acid tests, web developers don't need to work around the flaws in each different browser. We all benefit by getting web sites with fancy new features that work in all browsers. The scores are not meaningful, but are a way to motivate the developers of web browsers to fix their flaws so they're not embarrassed by a low or non-passing score.