devel > a couple of GIFTI questions
Showing 1-3 of 3 posts
May 1, 2008 04:05 PM | John Harwell - Washington University School of Medicine
a couple of GIFTI questions
I was contacted by a developer with some questions about GIFTI and
need some assistance with a couple of them.
1) Is it GIFTI or GIfTI? Does it matter?
2) Should we specify a format for the "Date" metadata element?
Perhaps similar to "15 Apr 2008 09:27:39" with the time being optional?
3) Are the metadata names case sensitive? I say no.
4) Can somebody provide an explanation of how UniqueID and SurfaceID are used?
Thanks.
John Harwell
1) Is it GIFTI or GIfTI? Does it matter?
2) Should we specify a format for the "Date" metadata element?
Perhaps similar to "15 Apr 2008 09:27:39" with the time being optional?
3) Are the metadata names case sensitive? I say no.
4) Can somebody provide an explanation of how UniqueID and SurfaceID are used?
Thanks.
John Harwell
May 8, 2008 03:05 PM | Ziad Saad
RE: a couple of GIFTI questions
1) I don't like the 'f' in GIfTI just like the 'f' in fMRI. So
GIFTI is my vote. The parent initiative's official name seems to be
NIfTI so we would be insubordinate if we went GIFTI.
Can anyone think of a reason why we can't just not care about it ?
2) I think a format for the date string is quite useful, the less the parsing options, the better. Your suggestion for a format looks good to me.
3) I leave this response to Rick
4) A UniqueID is used to uniquely identify a GIFTI data element or a surface. You can use the IDs to refer to element common to multiple GIFTI objects. For example, a bunch of surfaces sharing the same mesh do not need to keep a copy of the triangulation element, just a copy of its ID. You could use filename as references but those can be changed by users easily and a filename may point to a GIFTI object made out of many elements.
To my mind, SurfaceID is the same as UniqueID. Is SurfaceID defined or used somewhere in the current GIFTI specs?
cheers,
ziad
Can anyone think of a reason why we can't just not care about it ?
2) I think a format for the date string is quite useful, the less the parsing options, the better. Your suggestion for a format looks good to me.
3) I leave this response to Rick
4) A UniqueID is used to uniquely identify a GIFTI data element or a surface. You can use the IDs to refer to element common to multiple GIFTI objects. For example, a bunch of surfaces sharing the same mesh do not need to keep a copy of the triangulation element, just a copy of its ID. You could use filename as references but those can be changed by users easily and a filename may point to a GIFTI object made out of many elements.
To my mind, SurfaceID is the same as UniqueID. Is SurfaceID defined or used somewhere in the current GIFTI specs?
cheers,
ziad
May 8, 2008 04:05 PM | Richard Reynolds
RE: a couple of GIFTI questions
1. I agree, and think we should feel free to name the format what
we want. NIfTI is at least understandable, since NITI would be
the precise acronym, and the 'f' was apparently added to make it
sound better (unlike with FMRI). GIFTI gets my vote, too.
3. My vote here is for case sensitivity. Otherwise, in any instance
that they are handled computationally (via a program or script),
one must remember to treat those particular strings without case
(as opposed to all of the other strings in the file).
If we agree, for example, that Date should be the name of that
MetaData, then we can start using it (I see both AFNI and Caret
use 'date', while BrainVisa uses 'Date').
My general opinion is, whenever we want multiple software packages
to understand some piece of data, we should precisely define its
format. But that is just an opinion...
- rick
we want. NIfTI is at least understandable, since NITI would be
the precise acronym, and the 'f' was apparently added to make it
sound better (unlike with FMRI). GIFTI gets my vote, too.
3. My vote here is for case sensitivity. Otherwise, in any instance
that they are handled computationally (via a program or script),
one must remember to treat those particular strings without case
(as opposed to all of the other strings in the file).
If we agree, for example, that Date should be the name of that
MetaData, then we can start using it (I see both AFNI and Caret
use 'date', while BrainVisa uses 'Date').
My general opinion is, whenever we want multiple software packages
to understand some piece of data, we should precisely define its
format. But that is just an opinion...
- rick
