Richard Durbin visited LBL on Tuesday, March 29.
Those present included Frank Eeckman, Suzanna Lewis, Ed Theil, Eric
DeMund, Arun Aggarwal, Marat Boshernitsan, and John McCarthy.
This note summarizes our discussions:
1. CURRENT WORK AT CAMBRIDGE. Richard discussed work on ACEDB and
related software currently under way in Cambridge at the Sanger Centre.
[perhaps Richard and his group might want to write a brief summary of
this at some point for the newsgroup. My notes are rather sketchy, but
they clearly have a lot of interesting stuff going on. -JMc]
Ed noted that LBL is interested in trying out and possibly using many of
the analytic capabilties that Richard described. Arun will take the
lead role here for installing, testing, and explaining them to the rest
of the group at LBL.
2. ACE2 AND ACE3. Richard said that he basically likes most of the
things Jean has done with ACEDB 3.0. It just happened so quickly that
he was reluctant to go along with all the major model changes, etc. for
the whole worm community before giving it a bit more consideration and
testing. But their intention is to re-merge the two systems rather than
to continue divergent development. This will require some time and
effort, however, to get the two sets of code back into synch.
3. IMAGES. Frank Eeckman said that he and Cyrus Harmon had decided on a
single image file format (PICT) for the Mac version of ACEDB. This has
simplified things considerably. [Richard added the following:
We also discussed images in detail during the afternoon, and came up
with a new specification for the graphXxx calls that handle images,
which Cyrus and I feel we can both implement:
him using Mac PICT files and Friedemann/me using X images.]
4. SOURCE CODE FILE STRUCTURE. Cyrus and Frank have added new
directories
for mac code: w1mac, w2mac etc., without touching w1, w2 etc. They have
also done some restructuring within those files; Frank suggests
that we ought to consider adopting it for the standard distribution.
5. MAP DISPLAY. Richard outlined what he and Simon Kelly have been doing
to rewrite the map display modules, with emphasis on user-configurable
"columns" for different types of information. Arun Aggarwal explained
and demonstrated work he has done to separate out 3 different levels of
information in map displays: (1) entire display window; (2) one or more
maps; and (3) columns within individual maps. Two major goals of this
work are (a) capability to display multiple maps from different
organisms simultaneously (e.g., human and mouse, or different plant
species), and (b) capability to show simultaneously (side by side)
multiple (zoom) levels of detail for a particular region (e.g.,
cytogenetic, genetic, physical, P1 clone, sub-clones, and DNA sequence).
After some discussion, Richard concluded that the work Arun has done is
nicely complementary to what they have been doing in Cambridge -- and
that we should try to work together to develop a single map display
module. Arun will send Email to Simon and they will try to work out
details. Probably Simon will continue to work on column-level code,
while Arun will work on map and display levels.
Both of these will use the new map models Jean developed for ACEDB 3.0,
and the new extensions to support multi-species homologies (though these
may still require some further clarification and simplification).
The goal is to have a joint prototype running for the July workshop.
5. ?LONGTEXT. We briefly discussed the still unresolved issue of how to
deal with different types of text. Richard said that he and Jean had
tentatively agreed to make array class objects "second class citizens"
that must belong to some other B class object. If this were the case,
then Longtext would become essentially another (special) datatype. We
agreed that there needs to be a simple, consistent way to represent
beginning and ending of Long (i.e., formatted) text, which may contain
embedded spaces, carriage returns, etc.
We also agreed that how to represent LongText within the system is a
different (albeit closely related) question. For example, if Longtext
is simply part of a standard object, it could be inially "collapsed" and
could simply be expanded within the same tree display, or it could be
displayed in a separate window (as is done at present).
Richard pointed out that ACEDB does not currently have a multi-line text
edit capability. We noted that Gary Aochi had developed such a
capability last year, but it was not incorporated in the standard
distribution. Eric will try to revive that code and merge it with ACEDB
3.0, for Richard and Jean to consider for possible incorporation in a
future version of the standard distribution.
-JLMc
+=====================================================================+
| John L. McCarthy. . | Internet:..JLMcCarthy at lbl.gov |
| Computer Science R&D MS 50B-3238 | Bitnet: ...JLMCCARTHY at LBL |
| Lawrence Berkeley Laboratory .| telephone: (510) 486-5307 |
| 1 Cyclotron Road . | |
| BERKELEY, CA 94720, U.S.A. . | FAX: (510) 486-4004 |
+=====================================================================+
