|
SVN is not meant to move binaries, you would blow away their database
backbone with XSI project data. There is a good reason for Alienbrain
being a successful product despite open source version control systems
being available.
-André
Julien Stiegler wrote:
what's wrong with OSS solution such as Tortoise SVN client , outside
of xsi commit/update integration ?
I would rather look on the side of a python svn plugin integration for
update/commit automation.
Vincent Fortin wrote:
Is it reasonable to hope that the acquisition of Alienbrain could
complement XSI in that regard?
I agree with Raffaele that ref models should be developed with having
some kind of management system in mind. Or at least provide the
proper hooks for companies to build their own one on top. Personally
I don’t know the feasibility with the actual implementation and I’m
very curious to hear what issues others have encountered.
What I’d really like to see implemented alongside with the known bug
fixes are more flexible ref models, up-stream connections a.k.a more
stuff allowed in the Delta (deformers, topology changes, attached
properties and other stuff more related to fx). It seems that they
were mistakenly molded for character animation only and I’m afraid
that their usage is generally put aside because of that limitation.
Hopefully it’s a simple question of time; Delta is still very young
after all.
Thanks Kai for sharing your experience, keep up using them and report
your impressions to Softimage!
Vincent
------------------------------------------------------------------------
*From:* owner-xsi(at)Softimage.COM [mailto:owner-xsi(at)Softimage.COM] *On
Behalf Of *Steven Caron
*Sent:* Tuesday, February 19, 2008 9:20 AM
*To:* XSI(at)Softimage.COM
*Subject:* Re: experience with referenced models
"...but still It's pretty hard to completely separate referencing
from asset managent..."
agreed entirely... i just dont see what softimage could implement in
this area to help me without restricting me like project directories
do. everyone knows how much a PITA those can be.
On Feb 19, 2008 6:03 AM, Raffaele Fragapane
<raffsxsilist(at)googlemail.com <mailto:raffsxsilist(at)googlemail.com>>
wrote:
That's another subject entirely, but still It's pretty hard to
completely separate referencing from asset managent, and the two need
to have at least limited hook into each other, unless you work small
scale, where user focus doesn't waste enough hours for automation to
be worth the effort :)
On Feb 19, 2008 1:50 PM, Steven Caron <carons(at)gmail.com
<mailto:carons(at)gmail.com>> wrote:
it is sounding more like you are suggesting an asset management system.
but i hope it improves in ways you expect in the future.
steven
On Feb 19, 2008 5:21 AM, Raffaele Fragapane
<raffsxsilist(at)googlemail.com <mailto:raffsxsilist(at)googlemail.com>>
wrote:
Without droning on for hours about what is what...
A user centric system focuses on usability and being easily hackable
in and out of scenes, but also relies on the user to keep track of
things, pull data, version things and is oblivious to all of this,
focusing mostly on offering a solid empty shell with minimal file
monitoring and output.
A data centric system can be harder to hack around, is harder to
deploy (as it always needs a certain degree of customization), but it
relies on hard commited data, files, keys, QA and check processes to
guarantee, version and track data.
You can very easily imagine pros and cons of both.
My only real resentment with the current system is that it could have
done soooo much more with so very little additional effort. Offering
the option to customise keying and checks alone, and a very barebone
versioning backend would have gone a long way. As of now the only
thing that gets somewhat close are the onEvents, but they have always
been pretty dodgy to start with, and the way they are implemented in
such a blind way (they aren't away of anything and they aren't
designed as methods of a model) make them suffer from the same key
weakness in data awareness that XSI suffers for all over.
Basically something that could have been a big step forward in design
philosophy for the whole package has been slapped on with the same
exact strengths and defects the rest of the app has, which
incidentally is also what makes many people consider it a pipeline
weak software, and makes them prefer a stinking pile of **** like
Maya. It's a great chance to advance the application and show its
users new ways of work that, insofar, has been missed out on.
Does it make more sense now?
On Feb 19, 2008 1:04 PM, Steven Caron <carons(at)gmail.com
<mailto:carons(at)gmail.com>> wrote:
i think we have yet to see all that softimage has in store delta
referencing system.
you are right that the system is user-centric, but thats exactly what
i want... a consumable referencing system. now data-centric? i am not
sure what you mean. after all the user needs to manipulate the system
so that data can be stored then moved around the pipe in different
ways. they are bound by each other. again without specifics i can't
fully understand your angle on this.
you are also right about making certain decisions in beginning can
make things tough to change later down the road. what we have seen in
version 6 i believe is purely a gutting and replacing of the old
system. without insider knowledge on their roadmap we wont know
exactly what will be revealed to us in the next two versions.
i think what this conversation needs are some specifics or examples
or what we got is nothing concrete and nothing to request. with out
going into specifics we wont be able to fully appreciate your
experiences and how you imagine a better thought out system.
s
---
Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in body:
unsubscribe xsi
|
