-
Notifications
You must be signed in to change notification settings - Fork 4
/
105.text
193 lines (157 loc) · 8.8 KB
/
105.text
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
Subject: Progress report aips++ project, January 1992
Date: Tue, 11 Feb 92 16:30:55 EST
Please find below my report on the aips++ activities for the
first month of work.
Geoff Croes
Progress report on the aips++ project.
The aips++ project is, at the time of this writing, in its fourth
week. I will deal with the various areas in which work has been done
under the following headings:
organisation and facilities
user requirements
uvw data processing
image handling
user interfaces
basic libraries
system organisation and tools
intellectual property rights
CIC
software acquisitions
1. Organisation and facilities
The accomodation (apartments and rooms in Alden House) has worked
out well, as have the tiny 42 square feet offices. Due to some late and
unexpected arrivals (Maryland) we ran a bit short in everything, but
arrangements have been made to take care of everyone adequately. Every
participant has an IPX on his desk and there is a communal 486 PC
available for some special programs.
We had a C++ course in the week of 20 January. This was heavily
downgraded from what was the original intention to a medium difficulty
level. From what I felt to be the case from the "progress reports" on
training from the participants my expectations had dropped, hence the
downgrading. It turned out even worse than I expected: virtually none of
those who came here had analysed/designed/written a smallish problem as a
training exercise. As this constitutes the largest part of the half-year
ramp up to proficiency, much of that ramp up now needs to be done here in
the months ahead. This will affect progress and the result achieved by
July. The course was found too difficult by four of the about twenty
participants, far too easy by one and right-on by the others, a pretty
good batting average.
A consultant is in the process of being hired. He will join the
group for three days (Saturday, Sunday and Monday) each month. He has
degrees in Astronomy and Computer Science and seven years experience as a
leader of OO/C++ projects at Bell Labs.
2. User requirements.
We received a total of nine documents detailing user requiremnts.
These covered the requirements from the individual consortium members and
odd comments send in by users that had seen the NRAO document. These nine
documents have now been merged by Bob Hjellming into a single draft, full
of repetitions and waffly. Although we could circulate the document,
Bob prefers to do a bit more work on it before ciculation. It is, of
necessity, a document where wording is left almost untouched, which must
lead to large amounts of duplication.
The work groups, about which more below, have taken the original
documents and extracted those points relevant to their area, cutting out
duplication as much as possible. We do might merge the resulting
documents to provide an alternative merged user specification which is
closer to the work that will actually be done.
The work documents form the basis for the problem domain analysis
that is going on at this moment. The result of that analysis and the
initial design will result in the setting of priorities later this month.
Bob Hjellming and Brian Glendenning also spent a lot of time
trying to formulate a methodology for object oriented analysis with a
start in the astronomy domain. This methodology was discussed extensively
and will be the basis for the Green Bank and Charlottesville analysis
efforts. It maps pretty accurately into the methodolgies discussed in the
OO design/analysis books and the graphics CASE tools available for that
purpose.
3. uvw data processing.
A workgroup consisting of Bob Sault, Dave Shone, Mark Holdaway,
Chris Flatters and Sanjay Bhatagar has produced the uvw data processing
requirements report. This has been circulated and is being extended to
cover VLBI and single dish requirements in more detail. The group is now
in Green Bank, where they have joined seven domain experts to have a first
go at the analysis of this area. The others are: Tim Cornwell, Roger
Noble, Johan Hamaker, Rick Fisher, Phil Diamond, Bob Hjellming and Ed
Fomalont. The result of this meeting will be a report, edited by Tim
Cornwell and Dave Shone, expected by the end of February. The meeting is
chaired by Lloyd Higgs.
4. Image handling
This area is handled by a workgroup consisting of Lloyd Higgs,
Brian Glendenning, Peter Teuben and Mark Calabretta. User requirements
have been defined as a check list against entire systems that are being
considered for adoption. The group concentrates on defining a general
display model and of coordinate systems. It is also engaged in a thorough
examination of CIC.
5. User interfaces
This area is handled by Friso Olnon, Brian Glendenning and
Peter Teuben. The acivity in this area concentrated on an evaluation
of KHOROS, or at least that part of KHOROS that is of interest to the
aips++ project. Other studies deal with help files and history files.
They are still in the early stages.
222z
6. General libraries
222z
The foundation area of libraries for matrix and vector arithmetic,
container classes and portability interfaces is dealt with by Mark Stupar
and Bob Payne. They expect a first discussion document outlining the
contents of these classes and general implementation issues by the end of
this month. A discussion of graphics base classes has been postponed
until after a joint review with the image handling/user interface group has
been completed. The general feeling is that we might go with CIC and
Interviews, but that is a very early indication. The PEX library is also
being considered.
7. System configuration and tools
Mark Calabretta has written a survey of the system configuration
proposal for aips++. This has been ciculated and commented upon, and is
now being finalized. He is also working on a system installation proposal
and expects that to be available in another few weeks. The last part in
this series will be a proposal for networking, which he hopes to be able
to work on later during his stay in CV.
We are evaluating a number of release management systems to help
us in the development of aips++ code and documentation, and intend to
present a proposal for an acquisition at the March steering committee
meeting.
Chris Flatters has finished a proposal for source code
documentation. This has been circulated and a final version will be
distributed end February. We intend to adopt a book by Plum on C++
programming style as our de facto programming standard. Some work is going
on concerning a general standard for user documentation and help files.
We are inclined to go with Texinfo as our prime tool, but a definite
decision still needs to be taken.
8. Intellectual property rights
Our lawyers have advised us that the only protection of
intellectual property rights that is valid internationally is through
copyright. As we will have source code from many different contributors
(CIC, GNU, KHOROS come to mind, and there will certainly be others), the
only thing that seems to be reasonable to do is to allow different
copyrights clauses on different files, and arrange beforehand for
re-distribution licences, where needed. As a consequence, CIC could take
out its own copyright and licence the consortium members (but no one else)
to re-distribute CIC source code.
This leaves the protection on the aips++ source code the
consortium members write themselves. Copyright must be owned by someone.
There might be some problems in deciding, among the consortium mebers, who
that should be. There would not be great problems if we made the code pure
public domain. That would create the minimum of complications and
disputes. Opinions, please.
9. CIC
The representative for CIC (Chris Gunn) has arrived in CV. He
will provide various presentations to the aips++ group and participate in
discussions. Altough the first reaction to the CIC effort was favorable,
we have to ensure that CIC does fit well into a general, effective
display model. We expect a recommendation with regard to the adoption of
CIC as an integral part of the aips++ project by the end of February.
10. Software acquisitions.
The fact that aips++ is being developed by an international
consortium has made it possible to negotiate very good deals with
software vendors, who hope to get important exposure for their
product. We expect that the first deal that has been negotiated will
be signed this week. It allows members of the consortium to buy
PV-WAVE for a sixty percent discount on whatever their local price is.
We will be happy to provide details to those interested. NRAO intends
to buy for each site at least one copy that can be served to
individual workstations.
A second deal thast we are evaluating/negotiating is for a system
configuration management system TeamNet, that has been build around the
concept of distributed development. I expect to provide more details in my
next report.