forrest-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Robert Koberg <...@koberg.com>
Subject Re: Forrest HTML
Date Sun, 26 May 2002 14:35:02 GMT
Hi Christopher, thanks for the reasoned replies. (comments inline)

Christopher Bentley wrote:

> On Sunday, May 26, 2002, at 03:14  PM, Robert Koberg wrote:
>
>>>
>>> However, do we really need to know? Will browser and OS statistics
>>> actually help us? I think not. Let us stick to the cleanest
>>> (X)HTML code that we can manage with our powerful tools.
>>>
>> Well the cleanest would be using CSS :) But the reason to check what 
>> browsers the users actually use will tell us what browsers *need* to 
>> look perfect. Those that fall into a certain low percentage and 
>> display the page incorrectly are not considered in testing,  nor is 
>> time spent on trying to make it work.
>
>
> Hi Rob
>
> My Requirements Docs  used to say "Must render and function 
> Identically on IE5+ and Netscape 4.03+ on WIndow and Mac". They have 
> been increasingly relaxed about the 'render Identically' part for 
> about a year in regard to NS4.
>
> CSS is not as scary as it used to be, IE5mac, IE6 Windows and all 
> Mozilla variants have very nice solid support up through quite big 
> chucks of CSS2 and IE5 Windows although quirky and non-complient is 
> manageable in the most part and has pretty full CSS1.


CSS2 seems to be obsolete before it even gets going (CSS3 is well 
underway). I would defintitely stick with CSS1.

>
>
> I believe IE6 Windows use runs at around 60% in SME stats at present 
> with IE5 Windows taking up most of the remainder, as I mentioned in my 
> post "Cold sweat" your Users are likely to have scewed browser stats 
> and it would be good to find out at least what that scew is if in fact 
> it exists.


 yes, exactly. It could be that there are 30% of users coming in with 
Nav4.n. We should know to make sure.

>
>
> While the design I have seen will have no problem with a full on CSS2 
> base build, it would however place you at the bleeding edge with all 
> the pit falls of that place - would you be prepared to cope with the 
> problems (and possible delays) that may arise in the UI.. the 
> outstanding questions maybe.
>  *will it scale to met all the requirements to Phase 3
> * are there enough support people who have the CSS2 to maintain and 
> advance such a build
> * do you want to be an early adopter


On early adopter :) well... not for commercial stuff. For me, for 
commercial stuff, cocoon2 (production release) is still too young. 
Everything about cocoon seems like early adopter to me.

<fwiw role="religion">

But, taking cocoon's stated goal of SoC, why not extend that concept all 
the way down to the UA.

XSL           HTML structure (no presentation info)
+  ------->            + ------------------------------------------> browser
XML          CSS presentation

This would seem to be the most scalable and easily supported. You have a 
clean HTML structure separated from the final CSS presentation. The XSLT 
styling creating the HTML structure would rarely need to be changed 
going into the future. It is just SoC extended throughout the whole 
project. This seems to be right inline with cocoon ideals.

</fwiw>

Since you can't eat your ideals, I agree that full-blown CSS would not 
work, but it would surely be the cleanest, most elegant (code-wise) and 
easily supported (by the maintainers) way to go.

>
>
> I can't answer these questions, no one I know has ever implemented 
> such a build in a commercial production of any scale- we are all 
> waiting on Mozilla 1.0 to be released!.. but there has been some great 
> work done over the last year to develop strategies to to deal with its 
> arrival. There is a pent up desire both by clients and developers to 
> move to this kind of build - myself included. for the following reasons;
>
> The W3C recommend it
> It leads to accessibility best practices
> It is lightweight
> It is future proof (I'm sure we've all heard that one before but at 
> least IE 7 should'nt break it)
> It will work 'out the box' better against a wider cross section of devises
> It may even work ;)
>
> In the end Bert's view that he should center his design around 
> Netscape 4.08 may still remain the wisest for the obvious reasons not 
> the least because he is the one prepared to do it and is most 
> comfortable in that space. Also I sense a need for speed so adventures 
> in the unknown could be nerve racking for some

Sure, I think Bert would be providing an excellent service to the 
community if he can make elegant, easily maintainable HTML for the project.

>
>
> The downside to a NS 4 centered design is that it will not aid you in 
> keeping the UI lightweight, hack free or easily accessible.

Theres the rub. I have to use nested tables to get everything looking 
correct across browsers/platforms (even then I doubt you can cover 
everything). I can't make a clean HTML layout that can render in all 
browsers.

best,
-Rob

>
> But hey our Apache! at least think about outputing XHTML 1.0 
> Transitional... its XML!
>
> Chris.
>



Mime
View raw message