/[cvsroot]/skandha4/skandha4/To-do
ViewVC logotype

Contents of /skandha4/skandha4/To-do

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1.1.1.1 - (show annotations) (download) (vendor branch)
Tue May 25 09:39:23 1999 UTC (21 years, 5 months ago) by jsp
Branch: biostr, MAIN
CVS Tags: skandha4, skandha4_1_6, skandha4_1_6_pub, dsg_1_0, HEAD
Branch point for: skandha4_pub, autoconf
Changes since 1.1: +0 -0 lines
ANSI C version of Skandha4


1 Return comments:
2 ----------------
3 -> Hardwiring ../images/brainglobe.tiff pretty much
4 forces program to be run from given directory.
5
6 -> Having "init.lsp" hacked to run a separate demo
7 is a minor nuisance.
8
9
10
11 Date: Tue, 21 Mar 95 17:43:36 -0800
12 From: Jim Brinkley <brinkley@synapse>
13 Message-Id: <9503220143.AA06593@synapse.BioStr.Washington.EDU>
14 Errors-To: Jim Brinkley <brinkley@synapse>
15 Received: by NeXT.Mailer (1.100)
16 Received: by NeXT Mailer (1.100)
17 To: jsp@biostr
18 Subject: Possible skandha4 bugs
19 Cc: brinkley@synapse
20 Reply-To: brinkley@cs.washington.edu
21
22 Jeff,
23 I remember you asked me to mail you some of the possible bugs I've
24 found in skandha4. I've been keeping a list of bugs and questions which you
25 should have access to from betz.
26
27 The list of Bugs is ~brinkley/src/slisp/sk/Bugs
28
29 The bugs can be re-created by running some of the tutorial functions that I'm
30 working on, all of which are in ~brinkley/src/slisp/sk/lsp.
31
32 The idea that I've been working from is to define a set of files, each of which
33 shows progressively more complex functions along a particular dimension. Then a
34 higher level tutorials file picks from among these functions to build an
35 example. This should also make it relatively easy to allow these functions to
36 be chosen from the GUI.
37
38
39 The tutorial file is:
40 sk-tutorials.lsp
41
42 The comments at the beginning of this file list the other files that are
43 required. If you add ~brinkley/src/slisp/sk/lsp to *XLPATH* these files should
44 be loadable by just running one of the tutorial functions.
45
46 Many of the bugs refers to a function in sk-tutorials.lsp.
47
48 I'm still tinkering with all this, but I don't have as much time right now so
49 it won't change much.
50
51
52
53 Jim
54
55 ---------------------------Bugs-----------------------------
56 Possible Bugs
57 ----------
58
59 6. I have been trying to change the camera viewing transform in
60 various ways using the mouse. xg.3d-fly uses :turtle-roll. A function
61 I wrote uses :rotate-post. These work fine if used separately, but if
62 I want to move back and forth between these functions via keyboard
63 control I want the camera to be start out where it was when left by
64 the previous method. This doesn't happen when switching to
65 turtle-roll. The turtle roll message uses and updates the high level
66 fields, whereas the :rotate-post etc messages do not. So when you
67 switch from rotate-post to turtle roll the matrix is not consistent.
68
69
70 7. tutorial16 or 17 - I tried to let the user pick a cube, then move
71 it - using :down-click hook and :drag-hook. This sort of works as long
72 as the diameter of the camera is big enough. But once the cube gets a
73 certain distance away from the origin it snaps back to the origin. My
74 impression is that this may be the same thing as #6 - the :move
75 commands I'm sending to the matrix are not updating the high level
76 parameters, but at some point the system sets want_to_recompute_matrix
77 and uses the unchaged high level fields to re-compute it - which isn't
78 what you want.
79
80 Possible suggestion on transforms. I wonder if we need these high
81 level fields, at least the ones that affect the viewing
82 transform. Perhaps they could be embodied in a message to create a
83 transform given location, target and up, but these parameters would
84 not be instance variables. As far as I can see the only transform
85 messages that use these are the turtle messages, and I'm pretty sure
86 you can do the same thing by :rotate-pre messages - I did this and it
87 looks the same as the fly demo. I believe it makes sense to keep the
88 persepcitve transform fields like diameter and angle, since I don't
89 see anyplace where the transform messages affect these, other than
90 setting them directly.
91

brinkley@uw.edu
ViewVC Help
Powered by ViewVC 1.1.26