Thanks Raf,
This stuff is a big help. My biggest concern is with the fundamentals,
and your post confirms that this is the case, albeit with room for
improvement.
Send the teddy bear to Ubisoft Montreal, and dress him warm - winter is
coming.
Andre
-----Original Message-----
From: owner-xsi(at)Softimage.COM [mailto:owner-xsi(at)Softimage.COM] On Behalf
Of Raffaele Fragapane
Sent: November 5, 2006 6:59 PM
To: XSI(at)Softimage.COM
Subject: Re: Another Poll of Ref Models
Well, the subject is a bit hazy.
My comment on mixer inheritance and the positioning of data (which I
think is handled smartly enough) goes for models in general.
Going by the assumption that we're fine with the discrepancy between
facade and internal state (a hierarchy is presented when you really have
parallel name spaces under the hood), I think the way sources and clips
are positioned and used is logical, or at the very least convenient, as
it sweeps off right away issues that require tricky dependancy tracking
like multiple scopes, an override system etc.
I'm also fine with this being handled the same way for ref and local
models, as it enforces consistancy, and helps with referencing (having
something that can be safely isolated), having a different behaviour for
the two referencing modes (scene embedded VS external) would cause no
end of trouble, potential data corruption, and it would bring up even
more issues with conversion, that would have to either be left to
baffled users, or performed going by some heuristics that wouldn't
probably be smart enough to handle every possible case.
The thread is about ref models and how reliable they are these days
though, and going by that I think not everything is perfect yet, or at
least not in the form I would like it to be.
If you can understand this email you win a teddy bear.
******************************
| Raffaele Fragapane |
| Rising Sun Pictures |
| "Remember, TD is for TopDog" |
******************************
Kim Aldis wrote:
>You think the behaviour should be different in reference models to
loclised?
>
>
>
>>-----Original Message-----
>>From: owner-xsi(at)Softimage.COM [mailto:owner-xsi(at)Softimage.COM] On
>>Behalf Of Raffaele Fragapane
>>Sent: 03 November 2006 05:57
>>To: XSI(at)Softimage.COM
>>Subject: Re: Another Poll of Ref Models
>>
>>if we're talking about ref models, and not just localized models, Raf
>>doesn't think it's correct :)
>>
>> ******************************
>>| Raffaele Fragapane |
>>| Rising Sun Pictures |
>>| "Remember, TD is for TopDog" |
>> ******************************
>>
>>
>>
>>Kim Aldis wrote:
>>
>>
>>
>>>I've a few issues with them, none serious, so long as you know
they're
>>>there, just a bit annoying:
>>>
>>>Nested models:
>>>
>>> Mostly, more trouble than they're worth. I have problems with
>>>
>>>
>>some
>>
>>
>>>elements in nested models dropping out of groups and partitions. If
>>>
>>>
>>you
>>
>>
>>>select a root model and store an action, no actions on nested models
>>>are stored. Raf thinks this is correct, I'm inclined to disagree.
>>>
>>> Constraining reference roots to anything; they fall off
regularly
>>>
>>>
>>and
>>
>>
>>>xsi will crash a fair bit. It's more stable if you constrain the
>>>contents of the model, good enough reason in itself to build an srt
>>>root into the model.
>>>
>>> I've had problems changing some aspects of source models and the
>>>results of the changes not registering in references that are already
>>>in scenes and animating. Colours on materials, groups, and some
>>>
>>>
>>actions
>>
>>
>>>don't get picked up.
>>>
>>> Any nested models, the namespace isn't hierarchichal. You'll
find
>>>
>>>
>>the
>>
>>
>>>names of objects in the nested models will change if there's conflict
>>>with other models.
>>>
>>>
>>>I'm sure there's others, but these are the ones that spring to mind
>>>
>>>
>>>
>>>
>>>
>>>>-----Original Message-----
>>>>From: owner-xsi(at)Softimage.COM [mailto:owner-xsi(at)Softimage.COM] On
>>>>Behalf Of Andre DeAngelis
>>>>Sent: 26 October 2006 14:52
>>>>To: XSI(at)Softimage.COM
>>>>Subject: Another Poll of Ref Models
>>>>
>>>>Hello All,
>>>>
>>>>I am curious to see what the latest opinions were on the usability
>>>>
>>>>
>>and
>>
>>
>>>>stability of Ref Models. What people's experiences were with using
>>>>them and the dos and donts.
>>>>
>>>>We use them sparingly, but I am looking at how far to take their
>>>>implementation.
>>>>
>>>>All comments would be greatly appreciated.
>>>>
>>>>Andre
>>>>
>>>>---
>>>>Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in
>>>>body:
>>>>unsubscribe xsi
>>>>
>>>>
>>>>
>>>>
>>>---
>>>Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in
>>>
>>>
>>body:
>>
>>
>>>unsubscribe xsi
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>---
>>Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in
>>body:
>>unsubscribe xsi
>>
>>
>
>---
>Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in
body:
>unsubscribe xsi
>
>
>
>
>
---
Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in
body:
unsubscribe xsi
---
Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in body:
unsubscribe xsi
