- From: Addison Phillips [wM] <aphillips@webmethods.com>
- Date: Tue, 17 Jun 2003 17:44:34 -0700
- To: "Martin Duerst" <duerst@w3.org>, <public-i18n-ws@w3.org>
Thanks Martin. Added to the sign-up document. Others: where are yours? Addison > -----Original Message----- > From: public-i18n-ws-request@w3.org > [mailto:public-i18n-ws-request@w3.org]On Behalf Of Martin Duerst > Sent: Tuesday, June 17, 2003 2:37 PM > To: public-i18n-ws@w3.org > Subject: WS-Arch Usage Scenario S032: Caching > > > > I'm not sure I can do as good a job as Addison, but I'll give > it a try. > > Scenario S032 is about caching. The description says that this > can be on several levels, i.e. a whole SOAP message or just > some part. I'm not very sure the thing with the 'parts' is > practical, but anyway, this is about i18n :-). > > There are mainly two internationalization considerations: > > 1) The potential for caching gets reduced if the relevant messages > or message parts contain some i18n aspects, because only a subset > of the recipients will want to get exactly the same message. > This can be improved by using late localization as much as possible, > by including several alternatives (e.g. having a fault message > in 5 languages means that the same cached fault message can be > used for an audience covering these 5 languages, rather than just > one language). In order for caching to not produce errors with i18n, > the i18n context has to be well-known (explicit) to the caching > mechanism, or messages with i18n aspects have to be marked as non- > cachable. > > 2) In many areas of the world, connectivity is worse, and caching > therefore potentially more important, than in the better connected > part of the world. Example include Africa and some parts of Asia, > but also to some extent Australia/New Zeeland,... > > > Regards, Martin.
Received on Tuesday, 17 June 2003 20:44:43 UTC