This message is in MIME format. The first part should be readable text,
while the remaining parts are likely unreadable without MIME-aware tools.
--Boundary_(ID_nNUR74u4P3y72wNDAir9uA)
Content-type: TEXT/PLAIN; charset=windows-1252; format=flowed
Content-transfer-encoding: QUOTED-PRINTABLE
Dear Vera, Working group Colleagues!
D0 finished HEP authorship survey. It ended before December 1st 2005
deadline as we needed these data for our internal discussion of autho=
rship=20
rules. You can find summary of the D0 replies at the end of this mail=
. My=20
main comments based on the survey results as well as discussions at t=
he=20
different D0 bodies meetings:
1. Our working group survey questions are "poorely asked",=20
"leading"/"misleading" and "hard to interpret".
2. Most consider current HEP practice as "best choice". Small number =
of=20
respondents (only ~15 f the D0 Collaboration members participated a=
nd=20
~5 ad comments) is another indication that changes are not needed.
3. Partitioning of Collaboration is not favorable as well as importan=
ce of=20
including in the author list everybody participated in the hardware,=
=20
trigger, software developments is clearly stated.
In summary, attempting to "go back" to "a few" authors signing HEP pa=
pers=20
might not be productive. My recommendation is to accept changes in t=
he=20
way authorship is now treated in HEP and move forward by concentratin=
g on=20
explaining to the rest of scientific community advantages it provides=
,
continue developing mechanisms which identify particulary bright indi=
viduals
in the field and incensitives for individuals to "work hard".
Looking forward to summaries from other Collaborations as well as sum=
mary=20
for the full survey!
Dmitri.
*********************************************************************=
*********
D0 Survey Results
Number of respondents 93
Demographics:
Numbers are avg and number of respondents
Avg analysis ( 39.6 / 82
Avg detector ( 41.6 / 77
Avg accel ( 10.0 / 2
Avg manage ( 36.1 / 50
Avg age: 38.1 / 77
Gender: F 12 M 70 / 82
Survey Results:
Numbers are number in each bin and the average (93 respondents)
q1 CurrentPracticeBest=09[6, 5, 10, 10, 12, 18, 32] 1.14
q2 SciNotes=09=09[10, 13, 3, 12, 14, 21, 20] 0.61
q3 ContactPeople=09[33, 16, 9, 12, 9, 12, 2] -1.09
q4 BelleSignup=09=09[12, 12, 8, 8, 15, 23, 15] 0.41
q5 ReductEffect=09=09[10, 5, 16, 30, 7, 19, 6] 0.08=20
q6 MajorContrib=09=09[28, 15, 8, 4, 16, 16, 6] -0.60
q7 SmallResearchGrps=09[17, 16, 9, 6, 12, 20, 13] -0.01
q8 SmallResearchAuthors=09[27, 17, 8, 11, 12, 13, 5] -0.75
Comments (28 respondents)
user-comments: No offense but these questions are very poorly posed a=
nd=20
therefore any results of this survey will be suspect. For example, Q4=
asks if=20
the Belle practice is "a good way to identify and appropriately credi=
t" people.=20
I approve of the Belle practice but, since 50 f Belle signs the pap=
ers, I=20
don't believe it's a good way to allot credit. So I disagree with Q4 =
but I=20
agree with the Belle practice. I'm concerned that misleading conclusi=
ons will=20
be drawn from this survey.
user-comments: On Q2, that alredy exists in most of the collaboration=
s.
On Q3, the list of author changes with time. So this can lead to wron=
g=20
association of papers with authors.
On Q7 and Q8, it can be dangerous to have such a partition since some=
areas may=20
be overflown(for instace Higgs search) while others will have shortag=
e of=20
manpower.
user-comments: I think Q8 is very relevant. This would encourage peop=
le to=20
really spend some percentage of time doing analysis which is really w=
hat a=20
physcicist is here for, and not just doing software or hardware in or=
der to be=20
included in a publication. However, we have in HEP, many people who a=
re=20
non-physicists and just doing hardware or software. How they can be i=
ncluded in=20
an authorship in this scenario, is something to ponder upon. Perhaps =
a new=20
class of publications as addressed in Q2 would give credit to people =
not=20
involved in analysis, NIM publications etc.
user-comments: The current status quo has some drawbacks, but the add=
itional=20
burden and subjectiveness involved in selecting prominent authors is =
probably=20
deleterious to the overall physics mission of the collaborations.
Formalizing a partition of large collaborations into physics research=
groups at=20
the level of publication is harmful to encouraging research which cut=
s across=20
sometimes arbitray group structures and does nothing to acknowledge d=
etector=20
contributions. With many of our projects being decade-long efforts an=
d more, a=20
change to the system which does not reward (in fact discourages) dete=
ctor=20
contributions is very worrying for the health of the collective endea=
vor.
user-comments: Question 5 is only directed to scientists at the asst.=
prof.=20
level or above.
It does not raise the question of how a large alphabetic author list =
is=20
detrimental to junior scientists trying to get jobs. For junior scien=
tists in=20
HEP, the only metric on scientific achievement is word of mouth since=
the=20
current author lists give no indication of anything above nominal par=
ticipation=20
in a collaboration. It is clearly meaningless for someone with a few =
years of=20
experience as a scientist to have a publications list including hundr=
eds of=20
publications and this is clearly seen as meaningless to scientists ou=
tside HEP.=20
In all other scientific fields, the publications list is one of the m=
ain=20
metrics for scientific merit. In HEP, it is becoming more like a stam=
p=20
collection.
There is also the broader issue of the ethical misconduct associated =
with=20
claiming authorship for a result where the author=92s only claim to i=
nvolvement=20
is some fraction of their research time on some small portion of the =
experiment=20
and in many cases hasn=92t event read the manuscript. While this woul=
d be an=20
ethical problem in most fields, it is encouraged in HEP. In particula=
r, at=20
Dzero, scientists are discouraged from removing their names from publ=
ications=20
even if they do not feel they made a significant contribution to the =
science=20
presented in the manuscript.
user-comments: It is very hard to separate those who have worked prim=
arily on=20
the "service" tasks (trigger, reconstruction software, hardware, etc.=
), from=20
those who have concentrated on physics analysis. In any authorship pl=
an, all of=20
these (and other) contributions must be included in the author list. =
Without=20
that acknowledgement, the "service" tasks will not get done.
user-comments: I tend to have an alphabetical list of primary authors=
, with an=20
electronic link to the full author list/collaboration members. There =
is too=20
much battle already for the first-author position, we don't need more=
. This=20
practice accomodate the intention to have a list of primary=20
contributors/authors, and general contributors/collaboration members =
to the=20
publication at the same time.
user-comments: Much depends on the details of execution. I don't feel=
=20
particularly trusting that the field's hierarchy would produce a fair=
or=20
sensible execution plan for determining authorship in a different way=
.
user-comments: I find the proposals in this survey so immature as to =
raise=20
questions about their true intents. I can see where the suggestions a=
re coming=20
=66rom, and it is unfortunate that those who stand to lose are also t=
he ones that=20
are less likely to voice their opinions. How do propose to partition =
a large=20
collaboration into smaller physics groups? Even within a collaboratio=
n, where=20
it is imperative to do this for reasons of practicality, it is causin=
g all=20
sorts of problems. Many pieces of work, including analyses, span acro=
ss physics=20
groups. Under the proposed modifications, those that are skimming off=
the top=20
will exploit the rest even more.
user-comments:
My general attitudes in questions of authorship are the following:
1) As a field we already over-reward those who specialize in "physics=
" analysis=20
and under-reward contributions to "service", i.e., keeping experiment=
s running,=20
really understanding the data, improving reconstruction algorithms an=
d Monte=20
Carlos, etc. Young people notice this fact and it is very difficult t=
o get=20
enough people to volunteer to do service work as a result. Any change=
s to=20
authorship rules that would highlight the few people who made the fin=
al plots=20
or wrote the final words would be highly detrimental, in my view ther=
efore.=20
Such changes would exacerbate an already difficult situation.
2) Scientifically, I think it is important for all active collaborati=
on members=20
to take responsibility for all papers published by their collaboratio=
ns. The=20
sanction of taking one's name off of a particular paper is one that i=
s rarely=20
used, but it's very existence exerts, in my view, a powerful influenc=
e helping=20
to maintain scientific quality and integrity of publications within o=
ur field.
I need to expand on my answer to Q3, since I found this question to b=
e most=20
ambiguous. If the suggestion is for the "two contact persons" to be t=
he two=20
people most credited with the particular analysis, my answer would ha=
ve been=20
"strongly disapprove". I answered on the assumption that the "two con=
tact=20
persons" would be physics conveners and/or spokespersons and would th=
erefore=20
remain the same for periods of a year or two. That is, I appreciate t=
he desire=20
to save printed journal pages, but strongly disapprove of credit for =
particular=20
analyses to be given in print, since that violates the principle of g=
iving=20
adequate credit to those who keep experiments going and do all of the=
other=20
necessary "service" work.
I have to say that I found the wording of one or two these questions =
to be=20
somewhat "leading".
user-comments: Scientific notes sound in many ways to be the sorts of=
=20
publications that are already available to us but which are not frequ=
ently=20
taken advantage of. A publication that "details a specific analysis" =
sounds=20
like a PRD! Simulation publications can go in Computing in Physics Re=
search (or=20
something like that - I can't quite recall). Same for algorithms. If =
these=20
aren't "real" publications, how will they matter for, e.g., promotion=
and=20
tenure. Within the field may be one thing, but as soon as you go out =
of the=20
field, they would require endless explanation and would probably be d=
ismissed=20
by most people.
user-comments: Listing all members as authors does not make justice t=
o all=20
members. The number of articles signed by a member should reflect tha=
t member=20
participation and contribution to the whole effort. This way, a membe=
r would be=20
granted a certain number of papers according to the risponsibilities =
taken by=20
that member. The number of papers that would be granted to a member s=
hould be=20
previously defined and agreed when the member declares his/her dedica=
tion to=20
the collaboration. Some, of course, will sign all the papers, some wi=
ll sign 2=20
or 3 a year, others 4, 5, and so on. An authorlist with hundred of au=
thors (or=20
thousands as is expected for the LHC) becomes meaningless and ends up=
affecting=20
the evaluation physicists have from their institutions and funding ag=
encies.
user-comments: The current practice is best. For most of D0 RUN I, I =
had=20
significant responsibilities for the detector and software. This prec=
luded me=20
=66rom taking as big a role in physics analysis as I would have liked=
but I=20
rightly felt that the best way to contribute was by making sure that =
the=20
information being used by the analyses was the best it could be. Many=
of those=20
who were very active in the analysis were relative newcomers. That wa=
s=20
sensible, but one needed to give proper credit for those who nominall=
y did=20
"less" on the analyses. The only fair way is to inlcude everyone=20
alphabetically.
user-comments: If you only list people who do the physics analysis on=
the=20
paper, then who will waste their time on service work, such as: prepa=
re comom=20
sample data, hardware maintance?
user-comments: As a physicist if I spend 100 f my time for 10 years=
designing=20
and building important parts of detectors, and do not get time to do =
a=20
simulatoin or physics analysis, is my impact zero? this is insulting.=
People=20
would not be able to get their data, if not for my valuable contribut=
ion. Are=20
physicists only defined by the final analysis? In that case follow th=
e NASA=20
model. Make the data public to the rest of the world and let them ana=
lyze. Why=20
do I want only people in my collbaoration who can work out of root-tu=
ples,=20
making plots which a theorist suggested anyway. What an easy way to g=
et a=20
publication. Stand on the backs of people who make the detectors for =
you, rely=20
on the theorists to tell you how to extract the signal (which variabl=
es to=20
plot) and then you publish a paper in your name. That would ba a sham=
e for the=20
field.
user-comments: The person who made a histogram has not contributed "m=
ore" to a=20
physics result than the person who built the silicon tracking detecto=
r. They=20
all should be authors, on all papers, in alphabetical order.
user-comments: One option to resolve some of the HEP authorship issue=
s is to=20
repeat what has happened ~40 years ago when accelerator and experimen=
tal=20
physics teams "split". Have team (collaboration) which is=20
designing/building/operating detector and another team(s) (collaborat=
ion(s))=20
which are doing physics analysis. This is similar to some of astrophy=
sics=20
experiments as well. It might not be as easy split as between experim=
ental=20
physics and accelerator physics, but it would potentially create less=
problems=20
and be more efficient for the final physics results then deviding sin=
gle=20
collaboration into "consortia".
user-comments: The partitioning of a collaboration into smaller physi=
cs=20
research groups for attribution can only lead to political fiefdoms a=
nd in the=20
end only limit the amount of cooperation that can occur collaboration=
-wide.
user-comments: Another idea: Have 10 or so primary authors, in alphab=
etical=20
order, with two indicated as primary contacts. All others who worked =
on=20
components of the analysis get cited in movie-like credits (either at=
the=20
beginning or the end). For example:
Specific analysis code: John Doe, Kathy Smith
Software framework: Xu Calman, Susan Muson, Ben Tracker, ...
Detector operations: Kazu Monitor, Bill Runcoor, Keith Shifter, Betty=
=20
Triggermeistresse, ...
Detector construction and commissioning: Bob Builder, Q.T.O.Kay, Dan =
Solder,=20
=2E..
=2E.. and so on
user-comments: Questions are extremely leading, appearing to reflect =
the biases=20
of the survey authors.
user-comments: Everyone working on an LHC experiment is doing so beca=
use of=20
discovery potential. It is critical that those who do the work of mak=
ing the=20
data useable be on at least the same footing as those who did the fin=
al=20
analysis steps. Would you imagine the root jockeys would object if ce=
rtain=20
publications were strictly reserved for those doing "service" work?
user-comments: Not sure that establishing smaller physics research gr=
oups will=20
effectively reward people who spend most of their time on detector ha=
rdware and=20
software.
user-comments: The categories are missing something like "software=
=20
development/algorithms", therefore finally acknowledging the fact tha=
t there is=20
more to particle physics than hardware and analysis.
user-comments: Make the authors recorded in alphabetical order in a f=
ile=20
accessible electronically and make this gain in space in paper for mo=
re=20
scientific results. It is hard to exclude somone from a list, since s=
ome people=20
(for instance those who spend day and night in the control room tryin=
g to=20
improve data taking efficiency) are more useful than other but they d=
on't=20
necessarly contribute to papers directly (even some don't care at all=
).
user-comments: The current system is far from perfect. But the nature=
of HEP,=20
it's size, cost, time scales etc. makes it inevitable that there will=
be a lot=20
of specialization. If we are to get all of the work that must be done=
, done, it=20
is often impossible for people to work on all parts of the process.
Trying to define only that part of the process which results in the f=
inal=20
paper, the analysis, as being somehow "better" than the rest will and=
has only=20
created severe injustice, bad feelings and the loss of very good peop=
le from=20
the field.
ALL of these proposals try in one way or another to define the final =
analysis=20
stage as being somehow the only part that is "doing physics" and ther=
efore=20
worth recognition.
user-comments: In using the 'alternate scheme' some attention should =
be taken=20
to the possibility that a detector/physics consortium even with hard =
work of=20
their members may not achieve the results expected at the time of exp=
eriment=20
design due to unpredictable factors and might be strongly affected wi=
th respect=20
to other consortia in the collaboration.
user-comments: What does permanent untenured mean?
user-comments: I fear for the coherence of a collaboration when it is=
split up=20
as a consortium. Also "discovery" papers would not be signed by the w=
hole=20
collaboration, which I strongly disapprove: the work even in less tre=
ndy=20
working groups is necessary to achieve these discoveries. I think the=
principal=20
authors of a analysis get credit and recognition in our field by othe=
r means=20
than being first on an author list and we should avoid to loose time =
and energy=20
in determining who should be first on a list of authors. Not to menti=
on, that=20
also analysis results are most of the time due to a group effort and =
not a=20
single person and therefore iherently injustice would always be done.
To have only 2 contact person printed is valid in my point of view if=
these=20
contact persons are the spokes persons of an experiment.
I don't think the Belle initiative is going in the right direction: a=
=20
collaboration should approve its results in total and be responsible =
for them=20
in total.
--Boundary_(ID_nNUR74u4P3y72wNDAir9uA)--
|
icon
and there is no indentation. 
