Newsgroups: comp.software-eng,comp.sys.super,comp.theory,comp.dsp,sci.math,comp.org.acm,comp.org.ieee,comp.protocols.misc,comp.realtime,comp.arch,comp.databases,comp.lsi,comp.parallel.pvm,comp.parallel.mpi
From: robert@audioheads.com (robert bristow-johnson)
Subject: Re: Publishing Scholarly Work on the Web -- opinions?
Date: Thu, 19 Sep 1996 23:21:31 GMT
Message-ID: <1727@audioheads.com>


in case you've already seen this post, i'm reposting this 
because it appears our ISP dropped the ball on it.  when
there are multiple newsgroups, it sometimes doesn't do too
good.

In article <3236515D.5B58@cerfnet.com>, N. Gat <oksi@cerfnet.com> wrote: 
>
> Subject:  Publishing Scholarly Work on the Web -- opinions?
> 
> I'd appreciate any pro and con arguments on the subject of this letter.
> Since this matter touches upon all the sci.*, many comp.* and other
> newsgroups, I'll try to cross post this message.
> 
> The TechExpo Web site would like to open up a ScienceExpo section
> dedicated to the publication of scholarly works in HTML format (TechExpo
> already provides an array of services to the science & technology
> community; see http://www.techexpo.com).
>

...  <mostly very good stuff>
 
> 
> Obviously ScienceExpo will not referee papers; the authors' names and
> the institution they represent are put on the line -- so posting papers
> ON-LINE should be given as much or more care and thought as submitting a
> paper to prestigious journals
> 
>                         A FEW TECHNICALITIES:
> 
> (i) The entire paper must be submitted in HTML, and graphs, figures and
> charts in gif or jpg format  (this is a deviation from common formats
> requiring PostScript or TEX, to allow viewing papers using the
> newsreader built into popular browsers, and to allow indexing such
> papers by all the Web search engines)
> 

aren't there software tools that can convert PostScript into HTML?
if not, someone could make some dough creating such a tool.

...
> 
> Finally here is the most sticky point.  If the service if totally free
> of charge, I'm concerned there will be a deluge of frivolous
> publishing.  I consider a fee structure to act as a "potential
> Barrier."   A low, yet significant fee (hopefully) will discourage the
> unwarranted publications, but will not be too high to discourage worthy
> publications.

idiots with worthless "scholarly work" and a little money would
still publish. 

>  Perhaps academic institutions will receive a discount,
> but corporations will pay somewhat higher fee also to defray the costs
> of the service.
> 
> So one question is whether a fee imposed on publication will completely
> stifle interest?  Is for example, a one-time publication fee of $400 for
> a corporation, and $100 for academia too high a fee?
> 

i've always felt one of the most dishonest attributes of the
current scholarly publication system is that so many people
(perhaps most) publish (and submit for publication) not for the
reason that what they have to say is useful, true, and relevant to
the readers but that the real purpose is only to advance their
scholarly career (to get a PhD, academic job, research grant,
attention/respect/celebrity from peers, tenure, power, etc.).

this is reflected in the fact that many journals (e.g. some IEEE
transactions) charge a fee to the author to publish his/her
work.  this makes sense if the purpose of the journal is to
benefit the author (but then why should readers have to pay for a
subscription?  why should we even bother reading it?).

however, ostensibly, the reason for publication is to benefit
those who read the paper and gain from the supposedly new and
useful knowledge that the author has discovered.  if i have
discovered something useful and previously unknown, you should pay
me (or at least not charge me) for going through all the work
needed to polish the data and reasoning to make it "publishable"
and giving you that processed information thus saving you the work
of recreating and reprocessing the info or saving you from the
deprivation of doing without the information.

even though i do not expect scholarly journal authors to get paid for
their valuable work (they don't deserve to for the 90%+ of the
excrement that gets published these days), i _do_ expect and
demand that _good_ work not be penalized in any way.  we deserve to
see that work and the author deserves the credit for it and
charging the author discourages that from happenning.  in other
words, the work, by its merits itself, deserves publication or
rejection (or both, in controversial cases). 

a very good book exposing academic bullshit is "Imposters in the
Temple" by Martin Anderson.  check out the "Glass Bead Game"
chapter. 
 
In article <tazenk7fjfx.fsf@eng.sun.com>, chiba (keith.bierman@eng.sun.com) writes:
>gcouger@tacoma.ceatlabs.okstate.edu (COUGER GORDON) writes:
...
>> If you want to make it work I think it will need to be peer-reviewed and you
>> better get well respected ones for the first journal.

i agree.  perhaps there could be two levels of review.  first
reviewed, ever so briefly, just to ascertain that the submission
is remotely related to the topic of the group or "online journal"
that the work is submitted to.  the critiria is not that the
reviewer(s) think the paper is good or original or useful, just
that, if the group is about "Widget Science and/or Design", there's
something in there that looks like it could be about "widgets" and
it's not about Eastern European Poetry.  otherwize the submission
could be total bullshit (indicating the author knows nothing about
"widgets" and is an imposter) or a seminal gem that will be
referred to for 100s of years.  any readers could access the titles
and abstracts (in one web page) and/or the papers themselves (in
another web page if they want too) and comment non-anonymously if 
they choose too.  papers in this category live for a finite (2 
years, i dunno) amount of time and then, if they don't get
"promoted" to the "second level" get either deleted or 
archived in such a way that is extremely cheap and possibly 
inconvenient to get online. 

the second level of review is that some anonymous (to the author
and the public) reviewers say that the paper is good (i.e.
something new and valuable in it) and after any necessary
revisions, it gets moved into the "better" category and archived
forever and is always obtainable online.  now if some paper that is
"good" does not move into this category in some reasonable time,
there should be pressure from comment posters (who have read the
paper and believe it's good) to get the paper "promoted" to this
level.  this means two things: 1.) crappy papers that the
reviewers liked get promoted and 2.) crappy papers that were
popular with the constituency get promoted.  i think i could live
with that.  however, a paper that _both_ the reviewers _and_ the
constituency thought is excrement suffers an ignominious demise. 

>...
>
>Some fields are harder to get respect in than others. See the Sokal
>"hoax" 
>       http://weber.u.washington.edu/~jwalsh/sokal
>for an example.
>

it's a wonderful and badly needed expose'.  i hope it
contributes to some reform efforts.

>Some "respected" journals are no better than the internet already (I
>doubt that is the case for technical journals in the sciences).
>

it very well _is_ the case.  there is an awful lot of technical
self-serving bullshit out there. 

>I imagine that someday tenure committees and such may (or do)
>concentrate more on *references* to publications than publication
>itself. The real measure of a published work is the impact it has on
>others (of course, some clever groups of people just go around citing
>each other ... so one has to view such metrics with some judgement
>... not just as simple numbers ;>).
>

see "Imposters in the Temple" regarding that.

>For organizations which use such metrics, internet publication might
>be better than journal publication ... as the readership (potential
>people to be influenced) is likely to be larger than in many
>traditional journals.

i agree.

i dunno if the comp.dsp folk will like it or not, but i'm
adding that usenet group to the spam list (if you don't like
it, "well, excuuuuse me!").  comp.dsp people can find the 
original posts in comp.realtime and other newsgroups in the 
header with http://www.dejanews.com/ .

r b-j
wave mechanics, inc.
robert@audioheads.com    or    robert@wavemechanics.com



