David Crossley wrote:
>
> When i first discovered Cocoon about a year ago,
> i was amazed at the extremely poor performance of the
> xml.apache.org website. It seemed incongruous that
> Cocoon could be so amazing, yet the website was so
> slow. Then i realised that the site was not being served
> by Cocoon. So it must be the usability of the static HTML
> pages at fault.
Yes. It's mostly due to nested tables + the bunch of gifs from the
sidebar.
Anyway, over on [EMAIL PROTECTED] they are talking about the new
site look (and planning to use Cocoon2 to build it without us to say
anything! hurray :)
Maybe it's time to dive in and make xml.apache.org look as professional
as the software is hosts. (no offense to pier who designed the actual
site, but gosh, it's a dog!)
> We need to rectify this before release. The web pages
> loading needs to be snappy, to give the user a feeling of
> speed, power, and versatility that Cocoon can provide.
No doubt!
> Certain pages are a major problem. There are some
> bandages that we can apply to get us through the next
> release - i expect that a total rewrite of the document2html.xsl
> stylesheet is beyond scope for the upcoming release.
>
> Many pages are extremely slow to render, especially
> with Netscape. Perhaps this is because every page is
> constructed with complex nested tables.
>
> Two documents in particular make heavy use of images.
> tutorial.xml "How to create Web Applications" and
> concepts/index.html "Understanding Cocoon"
> If we use the height and width attributes for the figure elements
> then the browser can allocate the screen space and so can
> get on with rendering the text of the page. Then the images
> can download in their own time. I added patches today to
> fix this aspect.
>
> Does anyone see any other tweaks that can be applied
> to improve the website performance?
On option would be to add pagination.
For the image gallery I wrote such a transformer that uses information
in a "pagesheet" (sort of a stylesheet, but for pagination rules) to
filter out an incoming XML stream and create a page.
Unfortunately, it's *far* from being good enough for production :( [it
works perfectly for my image gallery which is mostly flat xml files, but
when I tried to test it with more nested and complex xml files, it
showed it's lack of proper design.
In short, I have to rewrite it almost from scratch :(
The gallery works but Laura (my girlfriend) is even more picky than I am
and doesn't want me to make the work public until it's "perfectly
finished". I know, I know, I tried to tell her about 'release early,
release often' but she think this might work for software since we care
for it, but it doesn't for usability since we are personally used to
un-usable stuff.
(yes, it took years to modify xml.apache.org and she almost puked on the
screen when she saw it.... go figure :)
So, result: I'll focus on writing the paginator module so that, at
least, I can make that public and we might try improving the side speed
with that by dividing the pages into smaller ones.
--
Stefano Mazzocchi One must still have chaos in oneself to be
able to give birth to a dancing star.
<[EMAIL PROTECTED]> Friedrich Nietzsche
--------------------------------------------------------------------
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, email: [EMAIL PROTECTED]
