Duncan,
I know open source development is different from commercial
development but if I could get a rough idea of how many man hours/days
went into developing the portal driver at least, it might save me from
_another_ impossible deadline.
Go on, make a wild guess, I won't hold you responsible for the
consequences ;-)
Man-years (of course) have gone into developing uPortal, an opensource
and relatively full-featured portal that uses Pluto for JSR-168 support.
http://www.uportal.org/
Why am I asking, well someone said to me the other day we (meaning me)
are going to develop our own portal and it needs to be done by Easter
2006 ...
A lot depends on what you'll use as starting material and what will
be your
requirements.
Remember you can use any of our codebases (pluto, jetspeed, etc...)
as a basis
for your own custom proprietary portal server (hell, IBM did it with
jetspeed 1
for their first Websphere Portal release so why not you ;).
That'll for sure save you a *LOT* of work.
Yes. I wouldn't presume to tell you what to do, but it seems pretty
clear that the sweet spot to be is to take an existing open source
portal (pluto, jetspeed, uPortal 2, uPortal 3, Liferay, ... there are so
many) and spend between now and Easter customizing it to accomplish
whatever is unique to your portal needs that was prompting you to want
to go write a new portal.
If you're looking to develop a custom portal you can then sell to
customers, most of these opensource portals have licenses that allow you
to do that. There are at least a couple successful commercial portal
offerings built on uPortal, e.g.
Andrew
