[thelist] WebReview responds to WaSP browser death march

aardvark roselli at earthlink.net
Sun Feb 25 23:18:20 CST 2001


> From: Andrew Forsberg <andrew at thepander.co.nz>
> 
> While I pretty much totally agree with everything that has been said
> about this (the webreview, Aardvark's comments, and the general
> feeling on the list towards this topic) -- I really want to play
> devil's advocate and ask a really basic question.

well, it's about time...

> 'Gracefully degrading' sites / sites that 'work in *all* browsers.'
> OK, so out with the details -- where is this holy grail of web
> development?

same place the real grail is...

> Sure divs just get rendered in the order they are written with no
> stylistic info on pre v4 browsers. So you can get the same content to
> everyone, but it looks like consecutive hunks of meat on v3 and
> earlier browsers. Is this really designing with pre v4 browsers in
> mind? or just letting them see something rather than nothing? This is
> largely an aesthetic issue however, since the content is separate, and
> the site is still accessible. With a modicum of care the whole thing
> will make sense on Netscape v2. I do think, though, that it reeks a
> little of 'you get the photocopy of the photocopy of the fax of an
> inkjet printer's output; and you, madam / sir, get the real McCoy.'
> And of course there's Navigator 4's 'implementation' of CSS and DOM.

that's part of the reason i have a problem with positioned divs... not 
enough people see layout... too many browsers see stacked 
content...  and i've seen plenty of ugly div-based code, as well as 
less-than-accessible content (thanks to positioning, invisibility, 
etc.)... bad code is bad code, no matter how you code it...

> On the other hand using tables for layout rather than purely 
> tabulated information is not a cool thing to do. Most sites do 
> though, don't they? These sites validate just fine with w3.org. The
> major problem, as I see it, with tables as a solution to reasonably
> compatible presentation problems is that they have a tendency to screw
> over users with disabilities' screen readers. (Or so I hear -- I have
> yet to play with one. Shame on me.) Oh, and tables are downright ugly
> from a code point of view (which seems to be the primary argument by
> Zeldman).

well, i still use tables for layout... are they strictly correct?  no, 
there is clearly a preference from the W3C for using tables only for 
tabular data... however, they recognize how tables *are* being 
used, and concede to that... they allow developers to do that by 
recommending ways to make them more accessible...

if you code them right, they'll degrade wonderfully in Lynx and 
others... and i've tested with alternative browsers... it can be just 
fine... *but* you have to know that up-front and code for that 
accessibility...

tables can be the worst thing you've ever seen (and usually are) if 
they aren't done well, but if they are, they can validate, still meet 
the spirit of the W3C and accessibility, and be used by anyone...

> It's either that or fork your code so that some get tables, others get
> divs. I'm not complaining about the extra work of forking a site --
> but it is a right nuisance to maintain on anything larger than a very
> small site.

ugh... nope, i won't develop two versions, never have had to... but i 
also haven't used all those barely-supported features of the latest 
standards just because i can...  doesn't mean i don't push the 
envelope, but my clients don't pay for it, so it has to be on my own 
time and on my personal stuff...

> This is a web design 101 type problem, so I'm quite reluctant to raise
> it here, but there is an important issue here that I think WaSP is
> making and not making, at the same time. Does that make sense? I mean:
> there is a problem with current website design and non-compliance with
> recommendations; but the problem is not older browsers vs newer
> browsers -- it is with developers writing good code (as Aardvark has
> pointed out). But, how 'good' is 'good code' -- I definitely end up
> simplifying sites I work on right down so that 9 times out of 10 they
> work surprisingly well on older browsers. On the other hand I don't
> even flinch at using validation breaking code in my header includes
> like:

i like to think that good code is tight, valid, does what you want 
wherever you want, and reads well... but others have different 
standards...

> <body bgcolor="#ffffff" marginwidth="0" marginheight="0" 
> topmargin="0" leftmargin="0">
> 
> Now. Hands up who hasn't ever used something like that! Huh? Huh?

actually, i'm doing my best to get away from that... but yeah, if 
that's all you have to do to get a site to work in NN4, then hell 
yeah, i'll add it...





More information about the thelist mailing list