On Thursday, September 18, 2003, at 11:30 PM, Stephen Colebourne wrote:
<snip>
Put it this way, I believe that j-c is creaking with too many projects and
changes are needed. But, I haven't yet seen any way that a-c is helping with
the problems. Instead it is adding new, nastier, ones. Thus I prefer to stay
with the least worst option of status quo.
it seems to me that the jakarta-commons was an experiment in the flattest of flat organizations. one mailing list for development, one mailing list for users, one web site, the only divisions being social ones. jakarta commons really is a bazaar, always busy, always active. this model appears to have been a spectacular success.
i suspect that other tlp's were inspired more by the organizational model (really, really flat) than by factoring our shared, reusable components (which is a major driving force for jakarta - and from what i can tell - httpd also). forcing smallish components together into a single flat community seems to work and that's what they are doing (so maybe xml-bazaar and ws-bazaar would be better names).
for example, i'd find it hard to argue that a JAXB implementation or an entity catalog resolver would be more in-scope here than in ws commons and xml commons respectively. when looked at in those terms, it seems to me unlikely these components would ever elect to move to here. so, it's really jakarta which is the issue.
stephen is right that the problem with this model is that the jakarta-commons cannot scale forever. we can't double the number of active components. we tried forcing httpclient out onto it's own list but this was IMHO a major mistake. maybe httpclient would benefit from a move to here where it could share a space with serf (lists, site and community). but they'd need persuading and i'm definitely not volunteering for that assignment :)
if this move was successful then i'd say that it'd increase the chances that other components in the jakarta commons would elect to move.
- robert
