llvm dev - Oct 2017 - Updating LLVM Publications Page

> On Oct 28, 2017, at 4:45 PM, John Regehr via llvm-dev <llvm-dev at
lists.llvm.org> wrote:
> 
>>> Incidentally, despite what that page says, 2009 was not the peak of
llvm-related-and-using publications.  It would be great for someone to do a
survey of papers out there and get more papers listed on the page. It would be a
great starter project for someone who was interested in contributing to llvm but
is just getting started in compilers...
>> Agreed.  I have a student that might be interested.  I'll check and
see.
> 
> Yes, let's put more thought/work into this. We need to accept that only
a tiny fraction of LLVM papers are going to get added by the paper's
authors. Searching for LLVM in Google Scholar seems to work well, and I have a
standing alert there to tell me about new LLVM papers.
We added the pubs page (and I force added news papers to it for a long time)
because there wasn’t a good external thing to point to.  Would it make sense to
“outsource” this to Google Scholar or some other service?

-Chris

On 10/29/17 11:30 AM, Chris Lattner via llvm-dev wrote:
>> On Oct 28, 2017, at 4:45 PM, John Regehr via llvm-dev <llvm-dev at
lists.llvm.org> wrote:
>>
>>>> Incidentally, despite what that page says, 2009 was not the
peak of llvm-related-and-using publications.  It would be great for someone to
do a survey of papers out there and get more papers listed on the page. It would
be a great starter project for someone who was interested in contributing to
llvm but is just getting started in compilers...
>>> Agreed.  I have a student that might be interested.  I'll check
and see.
>> Yes, let's put more thought/work into this. We need to accept that
only a tiny fraction of LLVM papers are going to get added by the paper's
authors. Searching for LLVM in Google Scholar seems to work well, and I have a
standing alert there to tell me about new LLVM papers.
> We added the pubs page (and I force added news papers to it for a long
time) because there wasn’t a good external thing to point to.  Would it make
sense to “outsource” this to Google Scholar or some other service?
I'm inclined to keep the publications page and to recruit new people to 
help keep it updated.  I think the publications page can help highlight 
to mainline LLVM developers how their work is used by the research 
community.  The page also provides a centralized place for LLVM users to 
advertise their research work that builds upon LLVM. It's mere existence 
points out that research is important to us.

On a related note, it's not important to me that every paper in the 
world using LLVM is listed on the page.  What interests me more is 
having the page advertise the fact that LLVM is used in a wide variety 
of research projects, including communities that have not traditionally 
used compiler techniques, and that it show a strong LLVM presence in the 
research community.  If we can get back into the habit of keeping it 
updated, I think it'll be a nice feature of the web page.

Regards,

John Criswell

>
> -Chris
>
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev

-- 
John Criswell
Assistant Professor
Department of Computer Science, University of Rochester
http://www.cs.rochester.edu/u/criswell

> On a related note, it's not important to me that every paper in the 
> world using LLVM is listed on the page.  What interests me more is 
> having the page advertise the fact that LLVM is used in a wide variety 
> of research projects, including communities that have not traditionally 
> used compiler techniques, and that it show a strong LLVM presence in the 
> research community.  If we can get back into the habit of keeping it 
> updated, I think it'll be a nice feature of the web page.
Maybe we can design the page using colors or something to help people 
see the different areas: security, formal verification, operating 
systems, HPC, embedded systems, etc.

John

> On Oct 29, 2017, at 9:04 AM, John Criswell <jtcriswel at gmail.com>
wrote:
> 
> On 10/29/17 11:30 AM, Chris Lattner via llvm-dev wrote:
>>> On Oct 28, 2017, at 4:45 PM, John Regehr via llvm-dev <llvm-dev
at lists.llvm.org> wrote:
>>> 
>>>>> Incidentally, despite what that page says, 2009 was not the
peak of llvm-related-and-using publications.  It would be great for someone to
do a survey of papers out there and get more papers listed on the page. It would
be a great starter project for someone who was interested in contributing to
llvm but is just getting started in compilers...
>>>> Agreed.  I have a student that might be interested.  I'll
check and see.
>>> Yes, let's put more thought/work into this. We need to accept
that only a tiny fraction of LLVM papers are going to get added by the
paper's authors. Searching for LLVM in Google Scholar seems to work well,
and I have a standing alert there to tell me about new LLVM papers.
>> We added the pubs page (and I force added news papers to it for a long
time) because there wasn’t a good external thing to point to.  Would it make
sense to “outsource” this to Google Scholar or some other service?
> 
> I'm inclined to keep the publications page and to recruit new people to
help keep it updated.  I think the publications page can help highlight to
mainline LLVM developers how their work is used by the research community.  The
page also provides a centralized place for LLVM users to advertise their
research work that builds upon LLVM. It's mere existence points out that
research is important to us.
> 
> On a related note, it's not important to me that every paper in the
world using LLVM is listed on the page.  What interests me more is having the
page advertise the fact that LLVM is used in a wide variety of research
projects, including communities that have not traditionally used compiler
techniques, and that it show a strong LLVM presence in the research community. 
If we can get back into the habit of keeping it updated, I think it'll be a
nice feature of the web page.
Makes sense to me, completely agreed.

+1 John’s down-thread suggestion of allowing a few tags for major areas.  This
could allow filtering on subject area and colorization.  Good luck defining the
categories though :-)

-Chris

On Sun, Oct 29, 2017 at 9:00 PM, Chris Lattner via llvm-dev <
llvm-dev at lists.llvm.org> wrote:
>
> > On Oct 28, 2017, at 4:45 PM, John Regehr via llvm-dev <
> llvm-dev at lists.llvm.org> wrote:
> >
> >>> Incidentally, despite what that page says, 2009 was not the
peak of
> llvm-related-and-using publications.  It would be great for someone to do a
> survey of papers out there and get more papers listed on the page. It would
> be a great starter project for someone who was interested in contributing
> to llvm but is just getting started in compilers...
> >> Agreed.  I have a student that might be interested.  I'll
check and see.
> >
> > Yes, let's put more thought/work into this. We need to accept that
only
> a tiny fraction of LLVM papers are going to get added by the paper's
> authors. Searching for LLVM in Google Scholar seems to work well, and I
> have a standing alert there to tell me about new LLVM papers.
>
> We added the pubs page (and I force added news papers to it for a long
> time) because there wasn’t a good external thing to point to.  Would it
> make sense to “outsource” this to Google Scholar or some other service?
>
One big obstacle to updating the publications page is that one has to
update a file inside an SVN repository. This might seem daunting to someone
whose only interest is to broadcast the presence of their paper. Even
writing to llvm-dev might feel like "too much effort". Can that page
be a
wiki instead?

Sameer.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20171030/e86b823a/attachment.html>