incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Endre StĂžlsvik <En...@stolsvik.com>
Subject Re: [Proposal] NoNameYet
Date Thu, 31 Jan 2008 14:57:25 GMT
kelvin goodson wrote:
> Hi all,
> 
> We've posted an Apache Incubator proposal onto the incubator wiki
> 
> http://wiki.apache.org/incubator/NoNameYetProposal
> 
> We haven't got a good name yet,  SandStorm is a contender, as is Snowdon
> 
> Suggestions and comments welcome,
> 
> Kelvin.
> 
> <http://wiki.apache.org/incubator/ThriftProposal>
> 

I have no say in this whatsoever, but I just find this a little funny:

   " Orphaned products

The contributors are leading vendors in this space. There is no risk of 
any of the usual warning signs of orphaned or abandoned code. "

Judging from IBM's donation of Pluto and WSRP4J, I do think there is 
quite some risk of huge commercial companies abandoning their code once 
they have rushed it through the JSR, or gotten some other means fulfilled.

The following points make this stand out even more:

   " Initial Source

The initial reference implementation code base is a donation from BEA 
Systems. The test suite for the TCK is a donation from IBM.  "

   " Homogenous developers

The small set of five initial developers are employed by two employers. 
... "


Pluto was a huge pile of dung when it hit Apache (I personally really 
mean that - the code was in my eyes unusable. This can probably also be 
evaluated against the adoption of that codebase, and stretching things a 
little, maybe even to the adoption of the whole JSR 168 portlets/portal 
scene, which I feel haven't gotten the following it clearly could have a 
potential of getting - look at MS SharePoint. I even believe IBM doesn't 
use Pluto for their own product WebSphere Portal, although I'm not quite 
certain of this. 
http://www.ibm.com/developerworks/forums/thread.jspa?messageID=14041773).

Pluto was pretty much left in that state for a long time, the original 
IBM developers of Pluto fleeing the scene days after the JSR got 
through. It sat there until someone else relatively recently picked it 
up and have now started to get it into some shape where it can actually 
be embedded into products as a the Portlet Container it was supposed to 
be. Compare that to the fact that JSR 286 is now nearly finished. Point 
2.16 of JSR 286 is answered as such "The RI will be implemented inside 
the open source project Pluto at Apache.". That project isn't exactly 
teeming with lots of different active developers.

I think Apache should be somewhat wary of this - don't let big companies 
abuse "the feather" as a playground, a rubber stamp of quality or 
something similar. Having a JSR's RI and/or TCK developed at Apache 
isn't worth much if it is next to unusable when finished. It could even 
be detrimental if the codebase is too bad: Apache can't start up another 
project or invite some other project since they already have the RI, but 
that is really so bad that no-one wants to actively develop it further. 
When something has been developed at Apache by these external companies, 
they should also pay the cost of getting a community up and fully 
running, so that at least the code can rather quickly mature into some 
usable version. Or potentially positively leave the RI "dead", starting 
or inviting some other implementation to be the "Apache Product".

Endre.


---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org


Mime
View raw message