Farewell, Darkstar, we hardly knew you

Not too surprising given that Jeff (the original developer) left Sun a year or so ago. However Jeff seems to be taking it back under his wing as an open source project, so I don't think it's fair to call it 'dead'. I actually have more confidence in it as a proper open source project than when it was controlled by Sun.

Unrelated directly to this project, in my irrelevant opinion, then next third-party "middleware" that will gain traction will be in form completely hosted and managed platform. If it hasn't already happened.

Those that have the capacity to set up their own data center will have the capacity to wire the "added-value" libraries in a way best suited for them.

This is based on observation of explosion of flash-centric sites, hosted web application development and deployment and closed gate communities aka XBLA/XNA, and partly Steam, Bnet, ...

A small project will meet the requirements without any forethought, using any technology. And any project that grows large enough will do so by being different enough to not fit into existing models.

A good, although somewhat slow adaptation to this model is happening in enterprise as well. On one hand, there is big iron. Very useful if one wants to have a large bully around. For everything else, the talent pool and availability of tools has grown large enough for cost of popular and free tools to outweigh cost savings of middleware.

Every middleware has non-zero learning curve, and majority of middleware at some point needs to be designed, or at least monitored by a "committee". This is consequence of funding - direction of development will be affected by those who pay for development (Linux kernel is example, big investors do not target desktop as a priority). So everyone who is not financing directly, is not in the "committee".

With barrier to entry for development being lower than ever, ingenious and differentiating products succeed, since mainstream is cornered by big bucks and existing monoliths. Middleware merely marginally reduces cost of mass production, something which doesn't earn money anymore.

This leaves room for other, boring details, such as billing, hosting, exposure, credibility/trust, ..., and that is where the "frameworks" would still add value, especially considering that these are the skills developers typically do not possess, or cannot afford to run themselves.

But as far as technical aspects go, they've been solved already in entirety, or they are solvable should the need arise. PHP/C++/Facebook and similar stories - by the time one encounters such problems, they'll be able to find people who can not only solve them, but much more importantly - can solve the right problems (one big issue with source code is that there are too many solutions looking for problems).

Quote:

Original post by Antheus Unrelated directly to this project, in my irrelevant opinion, then next third-party "middleware" that will gain traction will be in form completely hosted and managed platform. If it hasn't already happened.

That's probably the only economically viable option. Some people don't like it, but what can you do.


Brian Wood
http://webEbenezer.net
(651) 251-9384

"Unless the L-RD builds the house, they labor in vain that build it. Unless the L-RD watches the city, the watchman waketh but in vain." Psalm 127:1