Re: Warning about Netscape 7 Preview

>>You are correct that NN4 was released about 6 months before IE4, and that 
>>the HTML 4 and CSS standards had not yet been finalized.  So I do not 
>>fault NN4.0 for failing to support these standards.
>>I do, however, fault Netscape for not quickly addressing this issue:
>>the typical interval between major software releases is about 18 months, 
>>so I would have expected Netscape to have corrected this about a year 
>>after IE4 came out.  Instead (as I remember) Netscape issued NN4.5 
>>somewhat later, and its only significant concession to meeting standards 
>>better was to implement CSS1 more fully.
>
>That's news to me -- I didn't know that they added any more CSS support. 
>Except that I think the earliest versions only allowed one external 
>stylesheet to be linked in, while later ones allow more than one. Do you 
>happen to know what other things were added?

I know that in 4.5 Netscape (a) improved CSS support, (b) added user 
features, (c) improved stability, and (d) shoved down our throats a kitchen 
sink of non-browser features (AIM, etc.) which HAD to be installed and 
COULD NOT be un-installed, bloating it beyond belief.

As for CSS support, WebReview says that 4.0x supports 33% of CSS1, whereas 
4.5+ supports 76%.  The WebReview site probably has a table that itemizes 
the differences.  In any case, I test my sites using both 4.08 and 4.79, 
and it is easy to see where 4.79 does more than 4.08.

>>Of greater import were two other issues:  (a) NN4 violated standards by 
>>requiring that certain end tags appear in the code even when end tags 
>>were defined to be optional; and (b) when NN4 found something it did not 
>>like, valid or not, it tended to go berserk, either failing to render the 
>>page in a reasonable manner, or crashing, whereas when IE4 found 
>>something it did not like, it tried to recover gracefully.  These issues 
>>primarily affected designers who wrote invalid or sloppy code.  (I know 
>>that many designers think poorly of IE because its approach tended to 
>>hide coding errors; but I was always annoyed when Netscape crapped out on 
>>perfectly correct, 100% valid code.)
>
>(a) - The only case I know of that falls under this is Netscape's problem 
>with nested tables that don't have ending td/th/tr tags (for an example, 
>see http://wolves.dreamhost.com/web/html/tbl05.htm). I consider this a bug 
>rather than a standards violation.

Well, the standards are very clear that </td>, </tr>, etc. must be 
optional, and therefore that pages should be rendered correctly in their 
absence.  One could argue that what NN4 does is better -- many have made 
such an argument to me -- but regardless, the failure of NN4 to handle 
pages properly with such tags omitted IS a clear violation of the standards.

I should note, BTW, that it has long been my practice to use a code checker 
to ensure that I omit no end tags in my code, to avoid problems like this, 
and to avoid ambiguities which could result in unexpected rendering.  When 
xHTML appeared -- for which no end tag is optional -- I immediately 
switched to it, so now validation takes care of this issue.

>(b) - Well, crashing is certainly bad behavior.

I encountered a few CSS features that could crash NN4, and quickly learned 
not to use them.  Unfortunately I did not write them down:  I have a list 
of things in my head that I should not do in NN4.x, but no longer have a 
memory of why.  Sigh ...

>What I really wish is that each browser had the ability of producing a 
>nice report of all of the things in your code that it doesn't like. Sure, 
>we have validators that tell us how the code compares to the standard, but 
>I'd like to know whether there are things in the code that a particular 
>browser doesn't like.

The WebTV (MSN-TV) Viewer does something like this:  when it renders a 
page, it has a separate panel where it notes things it cannot handle.  And 
I think that iCab may do something like this ... maybe an iCab user could 
comment?

One of the things I like about Mozilla is that it can report a nice list of 
the JavaScript problems it finds:  much easier to debug JavaScript problems 
in Mozilla than in IE.


-
Chuck Upsdell
Email:     cupsdell(at)upsdell.com
Website:   http://www.upsdell.com/

HWG hwg-techniques mailing list archives, maintained by Webmasters @ IWA