[Pipet Devel] http protocol

J.W. Bizzaro bizzaro at geoserve.net
Sun Feb 27 12:14:11 EST 2000


Brad Chapman wrote:
> 
> > Here's a flowchart:
> [..snip..]
> 
> I *think* I am on the same page with you since my hand-drawn
> "Loci communications diagram" that I am trying to work off looks very
> similar to what you were diagramming, only a lot more ugly. We should
> try and make sure everyone is still in agreement on how things should
> go once we start to implement different protocols.

I'd like to fire up Dia to make some better-looking flowcharts.  And these
should go into the documentation Gary is preparing, which will be shown to
everyone here so that we can get their approval.

> For now, the
> front-to-middle communication is all I'm working on, and we'll have to
> worry about other stuff as it comes.

That's fine.  There's only so much a couple developers can do, and Loci is a
large project.

But I'm confident that once we make some usable releases, we'll catch a lot of
attention.  I've seen a number of projects on the net go from a trickle to a
waterfall of activity.  The most important thing in an open-source project,
IMHO, is that the developers stick to it.  As soon as project development
appears dead, the whole project and its use is truly dead.  People will move
on.

Speaking of usable releases, I think we should start planning for a 0.0.l
'alpha' release.  I know there is a lot of debate about what makes a program
alpha quality.  For example, Mozilla is only now considered alpha.  Developers
often use 'pre-alpha' to mean "don't even think of using this for anything".

I, on the other hand, think pre-alpha is the 'conceptual stage' where
relatively little to no code has been written.  Alpha is the 'initial coding
stage' where drastic changes can be made (perhaps we are already there).  Beta
is the 'testing stage'.  And after that you have your production releases.

So,  I think

  0.0.X  is alpha
  0.X    is beta
  X      is production

And pre-alpha has no release number.

Okay, probably not even half of open-source developers would agree.  But this
how I want to do it.

>     I just cvsed more of the changes towards complete separation
> between front and middle. The streaming dialog model is implemented
> for locus addition and connection, and all of the talking occurs
> through a port on the localhost, so from as far as I know security
> should not be a problem with this (unless someone already hacked your
> machine). There is a pretty weak communication log starting to be
> implemented, and the front now waits for a response from the middle,
> processes the response, and then does stuff with it (thanks for the
> Expect code--that was a big help!).

Wow, that is really great!

> This is just barely tested to
> work, and more will be coming (but after I spend some time with cactus
> and exams(ugh!)).

Cactus?

>     Hopefully all of my ugly code will be out of the front soon so
> that Jeff can work his magic on it!

Oh boy, oh boy!

> >Plea: Guys, for the sake of simplicity (and security!) can we just
> >deal with _one_ Internet connection (between middles)?  At least
> >until we've got this stuff down to a science.  What do you say?
> 
> Definately! 100% with you!

And you, Gary?

> >The 'INTERNET' connection is something we haven't discussed at all.
> >It involves 2 middles talking to each other.  Of course it can be
> >via standard Internet socket.  Some may argue for CORBA...I
> >wouldn't ;-)
> 
> I'm going to definately agrue for corba pretty heavily on this, but I
> think we should focus on local stuff first before we move into this.

Right.

> There definately isn't enough infrastructure in the middle to deal
> with middle to middle communications effectively now, so this is
> something I'll work on. Then we can fight it out over this....

I look forward to the battle ;-)

Cheers.
Jeff
-- 
                      +----------------------------------+
                      |           J.W. Bizzaro           |
                      |                                  |
                      | http://bioinformatics.org/~jeff/ |
                      |                                  |
                      |        BIOINFORMATICS.ORG        |
                      |           The Open Lab           |
                      |                                  |
                      |    http://bioinformatics.org/    |
                      +----------------------------------+




More information about the Pipet-Devel mailing list