Evolution of SIGSAM and ISSAC

[zimmerma]

Dear all,

I am clearly in favor of Option 3 (ISSAC leaves ACM) at https://sigsam.org/Evolution.html.
The reason is that I am against the "author payer" model that ACM moves into. This model
is what is used by the "predatory" journals (see https://beallslist.net/). It is not a good sign for ACM.

Paul Zimmermann

[MEngland]

I share your concerns about the "author pays" model. But I would note that the ACM's strategy and expectation is that most papers are covered by a "Read and Publish" agreement that is in place with the author's institution. It looks likely that over time most ISSAC authors would be covered by these. Of course, most does not mean all: if there are only a small amount of authors excluded from these then perhaps they could be covered by the ISSAC budget?

[Pierre Karpman]

I am mostly an outsider to this community, but I too would favour Option 3.
I am also wondering if it could make sense to join another “better” organisation, such as USENIX. I don't have any first-hand experience with it, but it has a clearly-stated and long-standing open-access policy (without APC) for the proceedings of its conferences. Admittedly it is probably a bit far thematically, though one could argue that as “the advanced computing systems association” it is not further removed from ISSAC than the “association for computing machinery”.

[rfateman]

SIGSAM problems are not just financial.

Executive Summary.
SIGSAM/ISSAC have changed. They concentrate on topics quite specialized
and rather different from their initial missions, and therefore
fail to appeal to their much larger potential audience.

To review briefly:

There are programs that are referred to as Computer Algebra Systems (CAS).
The current Wikipedia article
https://en.wikipedia.org/wiki/List_of_c ... ra_systems
lists 47 of them.

Some in that list are now obsolete, but a few have substantial user
communities. The builders and users of these systems have, in the
past, participated in SIGSAM.


How many users are there? Hundreds of thousands. Potential SIGSAM members.

More details:

Based on download statistics from SourceForge, the latest version of
the system I use most often (Maxima, a free descendant of Macsyma) has
been directly downloaded about 180,000 times between February and June
2024. Indirect downloads from other repositories for Unix systems are
likely. There are competitors, probably with thousands of
downloads.

Also there are too many new and diverse efforts in building free and open
new systems to neatly summarize them here.

There are much large user communities for other CAS, probably largest
Mathematica, Maple, Matlab but they are proprietary systems, and I
could not find a count of users. Perhaps others can add more
information on headcounts.

What can be used as a partial gauge of the community is the size of
the Mathematica stack-exchange site, with about 15,000 visitors, asking
and answering questions. The online Wolfram Community has over 37,000
registered users, and the Wolfram web site claims the company has
nearly 700 employees.

Some CAS are used routinely by large numbers of secondary
school and university students in mathematics and engineering classes.

What does this activity have to do with **today's** SIGSAM?

Very little.

The SIGSAM publication system has been incrementally diverted from its
initial domain of system building, algorithms and tools for use in
scientific computing: applied mathematics, engineering, physical
sciences, biology, etc.

It has been refocused as a target for publication of papers that
consist mostly of application of computers to tasks in pure
mathematics. These tasks are typically of very specialized interest,
in contrast to content of earlier years.

Most of the ISSAC papers are also too specialized to have broad appeal
in the world of symbolic computation.

While SIGACT may appear to the SIGSAM leadership to be the closest
other SIG to current SIGSAM activity, if there is to be a return
to the topic of building and using software, SIGACT is probably not
the place to sit. Although we could point to numerous SIG overlaps,
SIGPLAN and SIGNUM might be congenial partners.

Looking outside ACM, a match to interests of members may be with SIAM
SIAG/CSE. Decoding the acronyms: Society for Industrial and Applied
Mathematics, SIAM Activity Group on Computational Science and
Engineering.

For the computer-system savvy researcher it is easy enough to publish
a paper using Arxiv, or a source-code hosting site like Github or
SourceForge. These are not peer reviewed, but there may be other
mechanisms for collaborative quality filters. Consider the impact of
various discussion forums (compare to Yelp!) This is not the forum to
address these larger issues, except to note that the idea that authors
should pay for publication is problematical.

What is, for me, specifically a problem for SIGSAM and ISSAC: the lack of
relevance of papers that are published.

SIGSAM has lost its mojo. Can it be re-animated? If so, would this alleviate
the financial crunch? How might we best proceed?

Richard Fateman

(I was SIGSAM chair 1983-85. I'm an emeritus professor and former
chair of Computer Science at UC Berkeley.)

[afkafkafk13]

Pleaes allow me to contribute the point of view of Fachgruppe Computeralgebra, Germany.

The current discussion on the future of ISSAC and SIGSAM and the policy changes of ACM has spawned a discussion on this point among the members of the steering committee of the Fachgruppe Computeralgebra in Germany. We would like to summarize the resulting point of view as follows:

A) The status of 'sponsored by' has severe disadvantages in the context of the German law (similar observations may hold in some other countries, but we did not investigate this). In particular, such a conference cannot benefit from the facilities of the hosting university at the same, very favorable terms as a scientific event with the status 'in cooperation with'. These less favorable terms alone lead to an increase in costs and effort for the organizers of the conference, which in turn will necessitate an increase in registration fees. These drawbacks drastically decrease the probability of future bids from Germany, while, on the other hand, we all would like to continue the tradition of hosting ISSAC in Germany every once in a while.

B) Former organizers of ISSAC conferences in Germany (Munich 2010, Kaiserslautern 2017) pointed out that from their experience with ACM as ISSAC proceedings series publisher and with other publishers of other conference proceedings, they would not see a huge difference in quality and thus they would not see the difference in cost justified.

Overall, this leads the steering comittee of Fachgruppe to consider option (3) as the only one, in which from our point of view the advantages outweigh the disadvantages.

Anne Frühbis-Krüger
Sprecherin der Fachgruppe Computeralgebra, Germany

[alperenx]

I favor option two. My rationale is as follows: In conversations with other TCS experts I often notice there is a big lack of communication between computational algebra community and the broad theory of computing community. I think, as the current of state of the things are, it is hard to sustain a momentum in computer algebra without being integral part of communication in broad TCS community. So, I hope option can trigger new conversations/connections and help sustain our community.

Alperen Ergur,
UTSA

[JoeKileel]

Hi all,

I submitted to and published in ISSAC just once, but was impressed by the review process. From my perspective, the label of ACM is valuable to the computational algebra community. I agree with Alperen’s comments above: if anything, building bridges to other parts of ACM would be good. Therefore, my two cents are that options 2 or 1 are better, but not option 3.

Best,
Joe Kileel
UT Austin

[JohnA]

My preference definitely lies with Option 3 (ISSAC leaves ACM).

The advantage of Option 2 (merging) is presumably that the ACM annual fee
of $15k would be spread over a wider range of conferences & members (thus
becoming less per head), while ISSAC could "probably" continue much as
before (unless SIGACT decides to make changes).


I found Jonathan Aldrich's article a bit disappointing: it would
seem that he made little effort to contact the people running LIPIcs
to ask explicitly for information (e.g. when that information was
not found on the LIPIcs website). Nevertheless he did helpfully list
a number of features to take into consideration -- with weightings
appropriate to our case (and which likely differ from Aldrich's
priorities). Below are some of my doubts/differences of opinion:

(2) Maintenance
I would expect that name changes would be handled by ORCID numbers;
I am uneasy with the idea of retrospective revision of a published
article (as apparently implied by Aldrich).
As a digital format for articles (in Symbolic Computation) PDF is
not perfect; but I fail to see how supplanting PDF with HTML could
be considered useful progress beneficial to the Symbolic Computation
community.

(3) Digital library platform
Zentralblatt offers a good "digital library" service in the area
of Symbolic Computation.
Also I suspect that DOIs solve most problems regarding indexing and
citation-links (at least for modern articles with DOIs).
I'm quite skeptical of the genuine utility of "analytics"
(though some administrations may perhaps regard them as important?).

(4) Quality
Aldrich's comparison of ACM to arXiv is akin to "comparing apples
with oranges" -- we can dismiss it.

(5) Accountability
ACM handles "plagiarism & falsified data" in some way; LIPIcs might not.
How relevant is this to Symbolic Computation?

(6) Promotion
How important is "Clarivate" to the Symbolic Computation community?
Aldrich's comments suggest that analytics from Clarivate may be
"important" under some administrations -- does that justify imposing
APCs which are 10-20 times higher on everyone?

(7) Services
Aldrich claims ACM helps with reviewing while LIPIcs does not.
I am extremely skeptical of the first part of the claim; is there
any truth in it?

[JohnA]

A potentially useful piece of information:
Anne Frühbis-Krüger is the representative of Fachgruppe Computeralgebra which may be considered to represent essentially all researchers in Symbolic Computation in Germany (not sure if Austria is also included). So her one voice has many backers.

John Abbott (RPTU Kaiserslautern, Germany; formerly Genova, Italy)

[mourrain]

I am definitely in favor of option 3. Having (high) APC is not a good sign of quality and can generate issues in our community (requires resources, obstacle for publications of young researchers, ....). Having the stamp of ACM on the publications is not a sign of quality. The quality depends on the community, the chosen reviewers, their work, ....
Moreover, the argument of open access is a dummy argument, since one can publish our works on ArXiv or Hal (personally all my articles are available on HAL).
In some domains (e.g. math.), publications in ACM proceedings is not well considered (e.g. when applying for a job at the university) and publications in journals are preferred.
Option 3 is a great opportunity to have well-valued publications in a good journal (why not trying JSC ?).

Bernard Mourrain