[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: another application

From: Tom Lord <lord_at_regexps.com>
Date: 2002-10-13 17:41:48 CEST

> I think SVN will really gain popularity the day "public" svn
> repositories are available on the net.

That's a very hard goal. I don't think it should be a 1.0 goal to
prepare for that deployment. Here's some reasons why:

1) performance

   Big public repository hosts have controlled hardware costs and
   heavy work-loads. Such deployments are among the strictest
   performance requirements svn will _ever_ have to face. My strong
   intuition from reading the dev list is that the architecture is up
   to the challenge -- but the implementation will take longer. No
   surprise there at all -- it's essentially a really exotic file
   system implementation and those can take quite a while to properly
   tune.

   I say: go for a 1.0 that is usefully stable, then use the cachet
   of 1.0 adopters to develop commercial demand for tuning and
   scaling work. I think the wiki hack I described is an easy
   tactical play and fits the bill. I can even think of who might
   buy [*] a "development subscription" based on a 1.0 demo plus an
   outline of how the tuning/extension project would be run.

   [*] because it _really_ fits the business model of the company I
   have in mind, plus it fits the pattern of how they have in fact
   managed their costs and functionality over the years.

2) features

   The goal you named (deployment on big public project hosts) takes
   svn from a very small number of users to a huge number, almost
   overnight. I think work on the client will suffer from the mix of
   incentive-to-remain-upward-compatible plus
   overwhelming-number-of-complaints.

   (I've suggested elswhere to freeze the client, declare it a low
   level interface, and factor client work out to other projects
   post 1.0 --- I still suggest that, but it is orthogonal to the
   point that, regardless of the path the client takes, a first
   deployment to a large, demanding, and often rude peanut gallery
   is likely to introduce paralysis.)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sun Oct 13 17:40:08 2002

This is an archived mail posted to the Subversion Dev mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.