HEADS UP: Website Overhaul

Gary Thorpe gathorpe79 at yahoo.com
Wed Mar 10 13:32:36 PST 2004


d at xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Lines: 93
NNTP-Posting-Host: 149.99.114.171
X-Trace: 1078954337 crater_reader.dragonflybsd.org 185 149.99.114.171
Xref: crater_reader.dragonflybsd.org dragonfly.kernel:4208

Dave Cuthbert wrote:

> Gary Thorpe wrote:
> 
>> David Cuthbert wrote:
>>
>>> Personally, I limit myself to simple tables and font styles.  Oh, I
>>> know about the advantages of CSS and such; but preprocessed HTML works
>>> much more reliably.
>>
>>
>> Using tables to format text is bad style and the ugliest but most 
>> popular hack on the internet. It doesn't "look" that fancy when you 
>> have to use a text browser (instead of the memory gobbling mozilla).
> 
> 
> What should one use, then?  Frames?  Divs?  Neither of those look all 
> that great in lynx, either.


How about paragraphs and subheadings? Unless it is tabular data, it 
should not be tabularized. The only reason people do it is to get a 
magazine-like side list os links: its too bad a web page isn't a magazine.


> 
>> And if you seperate the style from the content, it will be easier to 
>> maintain: will it be easier to go through and edit a dozen documents 
>> that have particular fonts or edit the one stylesheet they all 
>> included for the same look and feel?
> 
> 
> Easier to keep the document separate from HTML completely and preprocess 
> it, bypassing buggy CSS implementations completely.


You are ignoring the time spent developing the scripts in PERL or 
whatever that will have to do this preprocessing as well as increased 
time to serve the page, increased server load/memory/disk/cpu, increased 
network load due to larger documents, and increased rendering time for 
the client (things like lynx can just ignore stylsheets taht are not 
embedded in the document an not even fetch them).

> 
>>> Of late, I've even foregone much of that, limiting myself to Wikis and,
>>> when a Wiki isn't available, wrapping everything in <pre> tags (see
>>> http://www.kanga-da.org/ for example).
>>
>>
>> This is dirt simple to do in standard HTML and you won't have to rely 
>> and manually inserting tabs/spaces: the layout cannot be that crucial 
>> anyway.
> 
> 
> Yeah, but manually inserting tabs/spaces becomes almost WYSIWYG in 
> emacs. :-)


WYSIWIG sucks. Thats why latex is still used widely for articles/papers 
and not MS-Word.

> 
> The layout is crucical.  It's impossible to read prose, especially when 
> interspersed with bits of code, when you don't have appropriate 
> paragraph breaks and indentation.  I got a readable layout on the first 
> try, rather than mucking around with HTML, saving, loading in three 
> different browsers to see what bugs the HTML exposes, editing, saving, ...


How about...using paragraphs, lists, blockquotes, etc? They are directly 
defined in HTML for your convenience.

You got a readable layout yes, but what happens when you add to it? And 
if you use normal HTML, it WILL render in almost any browser regardless. 
Oh, and trying to control how it renders is pointless as you have 
realized, so why bother trying to in the first place? And what is 
readable for you, may not be for everyone else, especially if they use 
some tool to automatically convert your page into another format (think 
text-to-speech). In these cases, a logical use of markup will probably 
be much better.

You think a table-ized, fonted page looks anything like you intend in 
lynx vs mozilla vs IE vs ....?

> 
> Nor do I have to worry about overly long lines.
> 
> And it's viewable in lynx, too!  Heck, it almost looks the same in lynx 
> as Mozilla!

Do you think you can try the many different web browsers to check how it 
"looks"?






More information about the Kernel mailing list