Doesn't having geographically located page caches reduce the doubling effect
in any given location?

Squids located in the US should be caching more en-US than en-GB, and those
in Europe should have more en-GB than en-US.

Jared

> -----Original Message-----
> From: [email protected] 
> [mailto:[email protected]] On Behalf Of 
> Gerard Meijssen
> Sent: 12 September 2009 09:48
> To: Wikimedia developers
> Subject: Re: [Wikitech-l] Language variants
> 
> Hoi,
> When we are to do this for English and have digitise and 
> digitize, we have to keep in mind that this ONLY deals with 
> issues that are differences between GB and US English. There 
> are other varieties of English that may make this more complicated.
> 
> Given the size of the GB and US populations it would split 
> the cache and effectively double the cache size. There are 
> more languages where this would provide serious benefits. I 
> can easily imagine that the German, Spanish and Portuguese 
> community would be interested.. Then there are many of the 
> "other" languages that may have an interest.. The first order 
> of business is not can it be done but who will implement and 
> maintain the language part of this.
> Thanks,
>      GerardM
> 
> 2009/9/12 Ilmari Karonen <[email protected]>
> 
> > Happy-melon wrote:
> > >> Ilmari Karonen wrote:
> > >>
> > >>> -{af: {{GFDL/af}}; als: {{GFDL/als}}; an: {{GFDL/an}}; ar: 
> > >>> {{GFDL/ar}};
> > >>> ast: {{GFDL/ast}}; be: {{GFDL/be}}; be-tarask: 
> {{GFDL/be-tarask}}; 
> > >>> <!-- ...and so on for about 70 more languages -->}-
> > >
> > > The above begs the question, of course, would this switch 
> actually work?
> > > And if it does, how does it affect the cache and 
> linktables?  More 
> > > investigation needed, methinks....
> >
> > Indeed, that was what I was wondering about too.  Without actually 
> > trying it out, my guess would be that it would indeed work, but at a
> > cost: it'd first parse all the 75 or so subtemplates and then throw 
> > all but one of them away.
> >
> > Of course, that's what one would have to do anyway, to get 
> full link 
> > table consistency.
> >
> > It does seem to me that it might not be *that* inefficient, 
> *if* the 
> > page were somehow cached in its pre-languageconverted state 
> but after 
> > the expensive template parsing has been done.  Does such a cache 
> > actually exist, or, if not, could one be added with reasonable ease?
> >
> > --
> > Ilmari Karonen
> >
> > _______________________________________________
> > Wikitech-l mailing list
> > [email protected]
> > https://lists.wikimedia.org/mailman/listinfo/wikitech-l
> >
> _______________________________________________
> Wikitech-l mailing list
> [email protected]
> https://lists.wikimedia.org/mailman/listinfo/wikitech-l


_______________________________________________
Wikitech-l mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

Reply via email to