Re: Re: New MCF pages

En réponse à mf <[email protected]>:
> What use is there in alt text for a logo ? A logo is something 
> graphical, and putting an alt text in there is only going to ruin the 
> layout in lynx. And it's not just tricking the validator, as I 
> absolutely don't care about being 100% HTML conform. Lynx displays 
> [image.gif] as alt text if no alt property is specified. Setting alt=""
> means lynx will not display anything, and make the page look cleaner. 

Yeah, that's what I said. It's a trick. Hopefully you're not making reusable
code in C or C++ here ;) Now I know Mozilla can render pages without loading
images. I wonder how it would render without them (never tested). Or maybe
download the website in my Palm to browse the spec whenever I want ;)

> The only image that makes sense to have an alternate text property has
> one, and that's the sourceforge logo.

Yeah, at least this is a consistent rule. Like in the image maps Tronic uses in
the spec, you can't just get out with a blank alt text.

> > Also why do you want to keep the space (%20) in the URL instead of
> using _ or no
> > space at all ?
> 
> Cause of the way the PHP script handles the menu.

I think I'll have to look at the PHP source to correct that. This is really not
nice/compliant. As I said, testing with Mozilla or MSIE is simply not enough.
MCF is meant to be usable everywhere (portable), the website should be the same.

BTW, people are pushing XHTML for various reasons, not just to annoy webmasters
(and force them to write clean pages). One of them is that handled device will
soon use XHTML-Basic (like cell phones). I think we should follow the standard
as close as possible, and I don't see any good reason to go into a non standard way.

> No, it just includes the page. You have to strip it yourself, but that's
> not going to be hard, is it ?

To edit a striped page, it will be. You have to add some the header/footer again
or remove it everytime you edit the page. At least if you want to do something
simple to update and you chose PHP, you could save some time/problems for the
person who is in charge of the content.

The best (fast) way would be to "pre-strip" the data after every page upload. It
can be done using a page.html.strip page next to page.html with the same time.
If the time of page.html has changed, page.html.strip has to be stripped on the fly.

> > Also, the title of the page remains the same wether it's on Home,
> License or
> > News. That's usually what is used as the link name in Google (or other
> engines).
> > 
> 
> No problem at all. Hikke could even write (he doesn't know but he could
> ;) ) a script that writes different pages for crawlers (these identify
> themselves with the userclient property).

Hum... So why normal people couldn't access these pages for web engines ? Do you
think it's safe to maintain 2 websites at the same time (even through scripting) ?

As for page stripping, you can also pre-process the site after every update, to
create the static pages once for all. That will save SF's CPU and make the same
site available to all. (only the dynamic parts would need to keep PHP)

I suspect that's how big news site are doing.

> > Don't get so nervous. I'm just trying to improve things. I've done the
> same
> > mistakes I'm pointing in the past, so I want to avoid it and share my
> knowledge.
> > Using 1 PHP page for a site is only good for security purposes in some
> cases,
> > not for a system that serves static pages (SSI was created long before
> to do that).
> Instead of including header and footer in every page, we now include the 
> page between the header and footer. I think it's easier that way.

I don't understand the difference.

cya


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf