Dear all,
In preparation of our TC6 meeting in Coimbra, here is the report
on the IFIP TC6 journal prepared by the TC6 committee on
"publication/QualityOfConference/LinkWithSpringer" set up in
Wroclaw.
We are looking forward to an interesting discussion in
Coimbra.
Best regards,
Guy
_______________________________________________________________________________
IFIP TC6 committee on
publication/QualityOfConference/LinkWithSpringer
Committee members:
R.
Boutaba
G.
Leduc, chair
G.
Pujolle
H.
Rudin
Topics to be addressed by the committee:
- IFIP TC6 Journal
- Quality of Conferences
- Link with Springer
Only the first issue (the IFIP TC6 Journal) is addressed here,
although some obvious link with Springer is also implicitly there. We
did not discuss the "Quality of Conferences" issue, but
noticed anyway that this topic seems to have moved to another TC6
committee chaired by Arun Iyengar.
Table of contents
- Open-access
- Electronic-only versus electronic+paper
- Copyrights
- Quality
- Reliable archives
- Web site
- Business model
- The journal
1. Open access
The journal should be open access
(http://en.wikipedia.org/wiki/Open_access)
2. Electronic-only versus electronic+paper
Electronic-only is enough. Anyway, if
needed an open-access electronic journal can be published
simultaneously as a hardcopy, either by an official publisher
or not.
3. Copyrights
Copyrights are retained by the author, i.e.
not even transfered to IFIP. This is the main asset of open
access: authors keep control on their work. Authors are likely to
appreciate this feature very much. 'No copyrights to IFIP' also means
'no problem with SSBM (Springer Science & Business Media)'.
We propose to replace the "copyrights transfer form" by
a "Code of ethics". This will solve the problem of verbatim
republication in absence of copyrights transfer. We would ask the
author to sign it instead of a copyright transfer form. We would
also require the author to use the URI of the paper (i.e. the link to
the journal DL) instead of putting the paper on his/her personal web
page (after all the journal DL would be freely accessible to
anyone).
If an IFIP conference paper (for which IFIP may still hold the
copyrights in some cases) is extended and published in the TC6
journal, the traditional problem of reuse of copyrighted material
applies. This problem will disappear when most IFIP TC6 conferences
will be published in the IFIP DL (and not any more by SSBM).
[Question] How can we use the IFIP name and logo for the journal
without having to apply to the usual IFIP bylaws (principally the
copyrights transfer).
4. Quality
Authors of best papers from our best conferences (Middleware,
WWW, Networking, ...) will be invited to submit extended paper
versions to the journal. We need a strong editorial board.
We should first provide and agree on a list of (broad)
topics and at least one expert per topic to be part of the editorial
board. An expert can be associated with several topics. We need to fix
a ratio between the number of editorial board members and the number
of "submitted+invited" papers: suppose we get 40 papers to
review per year (at the beginning at least), 20 editorial board
members are enough. This would lead to 2 papers per year per editorial
board member. Note that the editorial board is not a list of
reviewers, but a list of people responsible for handling the review
process, including the selection of reviewers.
The rule should be: only renowned and reliable people
should be in the editorial board. This list will set the tone of the
journal. No focus on one sub-community in particular (e.g. the
performance community). Fair representation of American, Asian and
European members. The purpose of this journal is not to set up an
"IFIP mafia" against other mafias such as SIGCOMM or
INFOCOM. If we are considered as another mafia, we will surely fail.
We should be open-minded.
For the open call, the papers should be submitted to the Editor
in Chief who will dispatch the submitted papers fairly to the
editorial board members.
[Variant] The IEEE Surveys and Tutorials uses an on-demand
reviewing procedure by the editorial board. Members are invited to
check the list of submitted papers and give reviewing preferences. We
could also implement that.
Authors will receive immediate
acknowledgment of submission and will be contacted within at most two
weeks by the responsible editor who will handle their
paper.
Short reviewing delays (3 months).
We can define a maximum number of very best papers per conference
(depending on the quality level of the conference) that are invited to
submit an extended version to the IFIP TC6 journal. Order of
magnitude: 2% of the submitted papers. For a big conference like
Networking, this would mean 9 papers.
[Question] Is there a problem with IEEE/IFIP conferences?
5. Reliable archives
Example of the open access 'Logical Methods
in Computer Science' Journal
(http://www.lmcs-online.org/index.php):
Disk archive and a hardcopy of the contents is maintained by the
Department of Theoretical Computer Science at the Technical University
of Braunschweig, Germany, and also by a large number of mirror sites
around the world. Something similar should be in place for the
TC6 journal.
6. Web site
We have three ways:
A. Experienced centre
We delegate this responsibility to a experienced centre,
which is already managing one or several (preferably open access)
journals. The cost, if reasonable, can be covered by IFIP TC6 for
example. A possible example is the GET in France which already
publishes "Annals of Telecommunications" and has some
resources to do so. There are other possibilities.
B. Google-like solution
Same as A, with Google handling the process. If the author
retains the copyrights (i.e. does not transfer the copyrights to IFIP
(and SSBM)), Google should not raise any copyrights issue here.
Probably no cost for TC6 either. I don't know if Google would accept
to handle the typesetting which is resource consuming.
Though less appealing than Google, there are other
possibilities, such as ISTE (International Society for Technology in
Education, http://www.iste.org/).
C. One of us is responsible for it with some support from his
institution or IFIP (?)
Free, open source software is available for those wishing
to start up new journals
For example the Open Journal Systems (OJS)
http://pkp.sfu.ca/ojs/ developed by the Public Knowledge Project
http://www.pkp.ubc.ca/index.html. While OJS is designed for academic
publishing, it can be used by anyone.
System Requirements for OJS:
To run OJS 2.x, your web server will need:
- PHP 4.2.x or later (including PHP 5.x) with MySQL or
PostgreSQL support
- A database server: MySQL 3.23 or later OR
PostgreSQL 7.1 or later
- UNIX-like OS recommended (such as Linux, FreeBSD,
Solaris, Mac OS X, etc.). OJS 2.0.2 and above supports Windows
servers.
Another open source software to run a conference or
journal:
- OpenConf:
http://www.zakongroup.com/technology/openconf.shtml
7. Business model
Depending on the solutions, the costs may differ. The following
costs are foreseen:
- typesetting costs (perhaps the highest of all costs
involved)
- managing the web site (storage of papers, interface and
reviewing process)
- ensuring the reliability of the archives
- possibly handling printed copies (perhaps only for reliable
archiving)
This is not a big deal for just a journal, but it may become
substantial if this grows to a full DL.
Perhaps an institution could handle it (e.g. The MIT manages
the 'Logical Methods in Computer Science'
Journal).
If Google (or ISTE) manages it, they get revenues from
advertisements (or use our journal as an advertisement for other
things).
If a university manages it, we don't know.
We need some continuity guarantee also. The journal should not
depend on the presence of an individual (e.g. a professor) at a given
institution. People move.
[Question] Could we estimate the typesetting costs? Guy Pujolle
will ask 'Annals of Telecom' to have estimates of the needed
resources/costs for this journal which publishes more or less the same
number of papers as our TC6 journal would. Note that IEEE charges $60
for the typesetting of a page (and the total cost for an extra page is
at least $200).
8. The journal
Name of the journal: IFIP Journal on Networking? [Is it
broad enough?]
[Variant]: IFIP Journal on Communications Systems [May be too
broad, we don't really address the physical layer in TC6]
Acronym: JONET? JoN?
Logo [Otto, any suggestion?]
ISSN Number needed.
Structure of the board (inspired by other Open Access
journals):
- Steering committee
It will be responsible for maintaining the philosophy of
the journal in the long run and to decide any changes in the direction
of the journal. For example, the steering committee should appoint the
editor-in-chief (EiC), then the associate EiC in consultation with
EiC. The steering committee has also to approve appointed area editors
and updates to the scope of the journal. The steering committee can be
composed of the founders of the journal. The EiC and other members
below should be renewed periodically. Maybe the steering committee
should also be renewed, e.g. (re)elected by TC6 on a regular
basis.
- One Editor-in-Chief (EiC)
- One Associate
Editor-in-Chief
He can handle special issues for example. He will
definitely have to handle all those papers for which the editor in
chief has a conflict of interest. The associate editor in chief should
also take over in case the editor in chief is
unavailable.
- Editorial Board
Small group of 20 people max. In case of conflict of
interest, the EiC may act as editorial board member.
- Liaison board
If the need arises, we can have liaisons to the following
bodies:
Liaison
to IFIP TC6, ACM SIGCOMM, IEEE ComSoc, IEEE Computer Society
Liaison
to conferences such as Networking, etc
Liaison
to some non open-access journal (e.g. IEEE, ACM, COMNET, ...)
Liaison
to W3C consortium
Liaison
to DBLP, http://www.informatik.uni-trier.de/~ley/db/
Liaison
to Google
Liaison
to University libraries
Liaison
to the directory of open access journals (http://www.doaj.org/)
Liaison
to CoRR: see
http://arxiv.org/corr/home
Journal of the Computing Research Repository
(CoRR)
Liaison
to the public knowledge project
http://pkp.sfu.ca/harvester/
...
- Web Master
- Periodicity, Call for papers and
Invitations to submit
We will start with a
quaterly-published Journal, which will carry full versions of the best
papers presented in the TC6 sponsored conferences. Assuming about 10
papers per issue, we believe that we will have a good supply of good
papers.
In the second year of its operation,
we will start accepting papers from the open literature that will
permit us to build it up to a monthly publication. The final goal is
to alternate issues, one with papers from TC6 conferences, and one
from the open call.
Several good quality IFIP TC6
conferences take place every year. The total number of submitted
papers is about 2000. The number of presented papers in a conference
that will be invited for publication in the new Journal will be
proportional to the number of submitted papers to the conference. In
this way, more papers from prestigious conferences (such as the
Networking series that typically attracts 450 submissions) will be
published in the Journal than from a lesser conference, which
typically does not attract many papers. Assuming that we accept 1
paper for every 50 submitted papers, we will have about 40 papers per
year. (These ratios can be changed depending on our
discussions).
The acceptance procedure will be as
follows. Once the selection of the papers has been concluded, we will
invite a number of authors to submit full papers. The papers will be
sent to referees along with the original reviews done for the
conference. The refereeing process should take at most 3
months.
Papers from the open call will
follow the normal reviewing process.
We need criteria to evaluate the quality of
a conference (our second mandate, not really addressed at this stage).
Perhaps some impact factor criterion. Anyway, if we apply the
principle above, namely inviting the top 2% of every TC6 conference
(where 2% is applied to the number of submitted papers), this
may be fair with respect to all TC6 WGs and events. By invitation we
mean 'invitation to submit an extended version which will be carefully
reviewed'. Workshops are not considered here.
--
________________________________________________________________________
Prof. Guy
Leduc Phone : +32 4 366
26 98
Université de
Liège Secr : +32 4 366 26 91
Réseaux
Informatiques Fax : +32 4 366 29 89
Research Unit in Networking
(RUN)
Email: Guy.Leduc@ulg.ac.be
EECS Department, Institut Montefiore, B 28, B-4000 LIEGE 1,
BELGIUM
http://www.run.montefiore.ulg.ac.be/People/GuyLeduc/