----- Original Message -----
From: Kevin A. Burton <[EMAIL PROTECTED]>
To: JetSpeed <[EMAIL PROTECTED]>
Sent: Thursday, February 03, 2000 12:23 AM
Subject: Re: New Portlet / Questions
> Josh Lucas wrote:
> >
> > I'm currently working on the Scripting News portlet and I'm making some
> > progress. For right now, I'm using Xalan to transform the XML to HTML
> > and return it to JetSpeed via the new ClearElement.
>
> Cool! I didn't think anyone saw ClearElement :)
I guess that's the benefit on pretty much every mailing list known to man...
:)
>
> > Originally, I wanted to use Cocoon for this but I also wanted to get
> > something up and running fairly quickly and I'm still learning the
> > Cocoon framework. I'm sure that porting to use Cocoon will happen soon.
>
> It should run under the CocoonPortlet IMO.
I think you are right which is another reason why I wanted to use Xalan for
the moment.
>
> > As I watched the performance hit because of the XML/XSL transformation,
> > some questions popped into my head.
> >
> > * How long does a url stay in the cache?
>
> Infinitly unless you provide a condition. Right now nothing is
> expired. However as soon as I get OCS support into JetSpeed then it
> will be an issue.
>
> See Cacheable.expire(). Overload this to return true when you want to
> expire your content. This way you can have it do things like check if a
> file has changed.
>
> For RSS I am going to have the Portlet wait a minimum of 1 hour before
> checking if it should expire itself. Then it will start checking it's
> PersistentDocument cached URL every 10 minutes. If it notices a change
> it will load the cached Portlet and then add a task to the Turbine
> scheduler to refresh the... remove it from cache and then populate a new
> one.
>
> > * If most channels are updated consistently, how will/does JetSpeed
> > handle this?
>
> I really don't expect most channels to be updated consistently. If you
> never want one to be cached you can override isCacheable to return
> false.
>
Cool.. That answers most of my cache questions. Thanks!
> > * Is there plans for some sort of 'harvesting' mechanism which will
> > bring down the content periodically, possibly pre-generate the HTML and
> > have it ready for JetSpeed on disk?
>
> Yes. I am going to write a "daemon" to run under turbine to fetch RSS
> and ScriptingNews docs from OCS (Open Content Syndication) streams.
> Thus the content will be updated on a regular basis. We should provide
> the same expiration mechanism BTW. Perhaps we should just write an
> object that can check for expiration and share the same object.
Hmmm, not sure how you does this in Turbine but it sounds cool.
>
> > For my job, I've written something which handles the last question.
>
> What does it do?
>
Basically, it's a Java app which gets called by cron every 20 minutes to go
out and 'get' the latest headlines from a variety of news providers which we
have for our(EarthLink) portal. The files are already in XML as they follow
a spec we gave them. Once we have the XML put into DOM, we send it through
an XSL Processor(currently Lotus but soon will be Xalan) to output for
various formats(HTML, JavaScript include file, WML, etc). That way we
already have the data and the portal is intelligent to know which file to
load based on the user's prefs. Does that make sense?
If you think something like this would be useful, I'd be more than willing
to re-write for JetSpeed (or Turbine).
josh
--
--------------------------------------------------------------
Please read the FAQ! <http://java.apache.org/faq/>
To subscribe: [EMAIL PROTECTED]
To unsubscribe: [EMAIL PROTECTED]
Archives and Other: <http://java.apache.org/main/mail.html>
Problems?: [EMAIL PROTECTED]
