[Cosmo-dev] Bug 10304 and base64 encoding
Matthew Eernisse
mde at osafoundation.org
Mon Sep 10 12:49:28 PDT 2007
Maybe I should have added that even long-term, I'm pretty strongly -1 on
anything that makes the UI even slower than it already is.
M.
Bobby Rullo wrote:
>
> On Sep 10, 2007, at 11:53 AM, Brian Moseley wrote:
>
>> On 9/10/07, Bobby Rullo <br at osafoundation.org> wrote:
>>
>>> Because we invented EIM+JSON.
>>
>> and it's documented on our wiki and available for anybody to write a
>> client to use.
>>
>
> I didn't say that no one COULD write a client for EIM+JSON, just that
> none exists. And it was created for the use of the web UI.
>
> I think other people using EIM-JSON is important, and I've already said
> that explicitly. I am just reminding you why it was created in the first
> place.
>
> However, I do believe that departures from standards are important for
> pragmatic reasons, as long as you offer the "strict" version by default.
>
> There is nothing that Base64 encoding helps from the point of view of
> the UI client, so I don't see the problem having it be optional. We've
> all talked about optimizations for the web UI that would result in
> departures from the Atom standard.
>
> There are performance issues in the Web UI, and it is documented (by me,
> on this list) that size of stuff (code + data) coming down over the wire
> is a big factor, so being stingy with bytes makes sense.
>
> Also, just because we don't (yet) have explicit performance goals
> doesn't mean we should never consider performance.
>
> To recap:
>
> Mandatory Base64 Encoding:
>
> Advantages:
> Standards compliant clients will get base64 encoded content
> Disadvantages
> Hard to debug
> Somewhat bigger downloads
> Decoding may be slow (TBD)
>
> Optional Base64 Encoding
>
> Advantages:
> Standards compliant clients will get base64 encoded content
> Disadvantages
> ...?
>
> Bobby
> _______________________________________________
> cosmo-dev mailing list
> cosmo-dev at lists.osafoundation.org
> http://lists.osafoundation.org/mailman/listinfo/cosmo-dev
>
More information about the cosmo-dev
mailing list