MozillaZine

Opera Now Most Compliant Browser, According to WebReview

Friday March 26th, 1999

CNet reports that WebReview's latest assessment of CSS1 compliance in release browsers puts Opera on top with 78%, followed by IE4 at 70.2% and Nav4 at 38.8%. Neither the M3 release of Mozilla or IE5 were assessed in the recent tally.

Contrary to the opinion of the study's author, Eric Meyer, who stated that Opera was "really very impressive, especially given its relative youth", a buggy 78% implementation hardly seems worth touting. As I've stated before on these pages, a browser that has is 78% compliant and buggy is as useful as a 10% compliant (but bug-free) browser. In fact, the 10% compliant browser is the better option, because no workaround code is required for implementation. However, the only browser that should garner any praise is the one that is 100% compliant, and bug-free in its implementation.

In addition, an interesting statement from the WSP that appeared in the CNet article is worth mentioning.

"The uneven deployment of CSS1 in major Web browsers over the last two years has caused Web authors great frustration and expense, and has won CSS an undeservedly obscure and difficult reputation"

From my experience watching the progress of Mozilla's CSS1 implementation, it seems that CSS's reputation for being difficult is deserved. Developers have had to search the CSS2 definition to clear up vagaries in the CSS1 definition. There is no reference CSS implementation *anywhere*, so developers are essentially guinea pigs and have to not only implement CSS1, but spend time and money working through all the inconsistencies that were not addressed by a reference implementation. It doesn't seem that the uneven deployment of CSS implementations has caused CSS1's reputation. If anything, CSS1's vagueness has had the direct result of limiting its adoption.


#13 Re:Opera Now Most Compliant Browser, According to

by mozineAdmin

Saturday March 27th, 1999 4:14 PM

You are replying to this message

Aleks,

I don't believe you got the gist of my argument. Maybe I could have been clearer. The point I was trying to make was this:

No matter how compliant a browser is, if what it implements is buggy, it's part of the problem, not part of the solution. NN4.0, 4.5, 4.51, IE 3.0, 4.0, 5.0, Opera 3.0, are all part of the problem, not part of the solution, because what they do implement is buggy.

First, you state "...then you say that somethine MORE copliant is WORSE than something LESS compliant, whether it's Opera or IE"

I did not say that. I said that an 80% buggy implementation is worse than a 10% proper implementation. There's a difference between that and "more is worse than less".

Next, you stated, "You're just trying to make excuses for Netscape because they happen to be less compliant (and for good reason, their browser was released much earlier and supported specs that were final at the time)."

I did *no such thing*. I have never excused Netscape's lack of compliance. In opinion pieces from a few months ago, I tried to cut them some slack when everyone was pounding them viciously, because we had Mozilla builds that could verify their progress (which is clear, and impressive). However, I *never* made any statements trying to excuse Netscape's lack of compliance in their 4.x implememtations. If you think that's the case, please feel free to find any statements from me that support your contention.

Also, you stated, "There are good excuses for Netscape's lack of compliance, and it's true that IE5 could be MUCH more compliant than it is, but saying that MORE is LESS is pure bullshit, and you have to be really creative to find reasons to justify that."

I don't think you have to be creative at all. The fact is that a 10% compliant browser that handles that 10% perfectly *is* better than an 80% compliant browser that has a buggy implementation. Why is this? Because when you code for that 10% compliant browser, you have to use no workarounds and you won't have to rewrite that code later. You can't use all of the spec, but the code you do use will work in that browser and every other compliant browser that comes after it. When you code for the buggy 80% browser, you won't get the expected result in other compliant browsers that implement the spec properly, and when that 80% browser is fixed you will have to recode your pages again. See NN 4.0's DOM implementation (which came out before there was any finalized spec) and IE 5.0's DOM implementation (which is still wrong a year after the DOM spec is finalized). Both end up being curiosities more than useful implementations, because everything coded for them will have to be recoded for truly compliant implementations.

Finally, you stated that, "99% CCS2 compliance is worse than 50% CCS2 compliance according to you!"

This is a misunderstanding of my argument. If that if that 99% compliant browser implements the spec perfectly, then it is indeed the better browser. However, if it implements the spec in a half-assed, buggy way, and that 50% browser implements its 50% perfectly, that 50% browser is the better choice, IMO - for the reasons I stated above.

I look forward to your explanations as to why a buggy implementation is better than a compliant implementation of an existing standard. *That* should be creative! :)

On a side note to Jonathan, who made a comment above yours:

I'd prefer to give no points to any of the past and present release browsers, because they're incomplete and buggy in their implementations of the standards. This goes for Netscape, MS, and Opera (and any other browser you can think of). Netscape got *wholloped* when people found out that the Mozilla 5.0 would not be CSS1 complete (before they moved to the new layout engine). I defended it on the basis that shifting to a new engine when a new release was about 3/4 of the way done was a bad financial decision, when work on the new layout engine would be continuing (and verifiable). Now that they have made the move to the new layout engine, any Mozilla bashing is really in bad taste, IMO. They've moved their release schedule by 9 months to a year to do the right thing.

If Opera's decision to release a non-compliant browser was a purely financial one, then I would give them the benefit of the doubt. I did for Netscape, and am willing to extend that kind of consideration to them.