Need GSoC mentors and help updating ideas page
383 views
Skip to first unread message
Aaron Meurer
Jan 24, 2024, 1:49:53 PM1/24/24
to sympy
It is time again for us to apply to Google to be a GSoC mentoring org.
I need help updating the GSoC ideas page.
https://github.com/sympy/sympy/wiki/GSoC-Ideas. If any ideas there are
already completed, please remove or update them. Also please add new
ideas, although you may want to discuss them here first if you aren't
sure about them.
Secondly, we need to update the potential mentors page.
https://github.com/sympy/sympy/wiki/GSoC-Ideas#potential-mentors. If
you were a student in the past I encourage you to mentor. Please add
your name there if you can mentor this year. If you can't mentor but
want to help review applications, just let me know.
If you are someone who is interested in applying to SymPy for GSoC,
start here. https://github.com/sympy/sympy/wiki/GSoC-Student-Instructions
Aaron Meurer
I need help updating the GSoC ideas page.
https://github.com/sympy/sympy/wiki/GSoC-Ideas. If any ideas there are
already completed, please remove or update them. Also please add new
ideas, although you may want to discuss them here first if you aren't
sure about them.
Secondly, we need to update the potential mentors page.
https://github.com/sympy/sympy/wiki/GSoC-Ideas#potential-mentors. If
you were a student in the past I encourage you to mentor. Please add
your name there if you can mentor this year. If you can't mentor but
want to help review applications, just let me know.
If you are someone who is interested in applying to SymPy for GSoC,
start here. https://github.com/sympy/sympy/wiki/GSoC-Student-Instructions
Aaron Meurer
Oscar Benjamin
Jan 24, 2024, 2:54:44 PM1/24/24
to sy...@googlegroups.com
Hi Aaron,
I think we should start by removing everything from the ideas page and
moving it to an "old ideas" page with a clear note indicating that
some might still be relevant but others are likely not. I don't think
it is good to leave things on there and just let them roll over from
one year to another.
Each idea should be added each year by someone who is at willing to
mentor the project in that specific year. Applicants can also look at
the "old ideas" if they are considering proposing something that is
not listed in the main ideas page.
It would be better to have a small list of good, still relevant ideas
that we are confident can be mentored and that have been recently
checked to see that they are not out of date. I don't think that we
can get to that without first just removing everything. If anyone
wants to see something on the "official" ideas page then they should
add it and list their name with a clear understanding that they are
willing to mentor that project this year if there is a qualified
candidate for the project.
Many of the names listed as potential mentors are people I have not
seen for some time. I don't really expect that they would mentor a
project and I also don't expect that they would remove their name from
the list: we should not carry that list of "potential mentors" over
from one year to another. The same considerations apply to many of the
project ideas that were added a long time ago by people who are no
longer around and likely have no intention of mentoring any projects
or updating the ideas page.
Oscar
> --
> You received this message because you are subscribed to the Google Groups "sympy" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to sympy+un...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/CAKgW%3D6JObtA47p10wX6mb5W6c3p%2BqRRQY4GBLnXHKZmyNUtNGg%40mail.gmail.com.
I think we should start by removing everything from the ideas page and
moving it to an "old ideas" page with a clear note indicating that
some might still be relevant but others are likely not. I don't think
it is good to leave things on there and just let them roll over from
one year to another.
Each idea should be added each year by someone who is at willing to
mentor the project in that specific year. Applicants can also look at
the "old ideas" if they are considering proposing something that is
not listed in the main ideas page.
It would be better to have a small list of good, still relevant ideas
that we are confident can be mentored and that have been recently
checked to see that they are not out of date. I don't think that we
can get to that without first just removing everything. If anyone
wants to see something on the "official" ideas page then they should
add it and list their name with a clear understanding that they are
willing to mentor that project this year if there is a qualified
candidate for the project.
Many of the names listed as potential mentors are people I have not
seen for some time. I don't really expect that they would mentor a
project and I also don't expect that they would remove their name from
the list: we should not carry that list of "potential mentors" over
from one year to another. The same considerations apply to many of the
project ideas that were added a long time ago by people who are no
longer around and likely have no intention of mentoring any projects
or updating the ideas page.
Oscar
> You received this message because you are subscribed to the Google Groups "sympy" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to sympy+un...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/CAKgW%3D6JObtA47p10wX6mb5W6c3p%2BqRRQY4GBLnXHKZmyNUtNGg%40mail.gmail.com.
Aaron Meurer
Jan 24, 2024, 5:40:56 PM1/24/24
to sy...@googlegroups.com
Hi Oscar.
I agree with all of this in principle, but my experience is that it's
hard to get people to help out with updating the ideas page. And even
if I wanted to do it all myself, I don't know enough about some of the
ideas (like the ones in the physics submodule) to say anything about
them. I will say for a lot of the ideas I do know about, they are
actually still relevant, the issue is just that they are very
difficult so no one has ever been able to do them. It is worth keeping
them around in case someone actually comes along who can actually
manage them.
If you want to take a shot at doing this, you have my blessing. I
don't think it's necessary to move the current page to an "old" page
since the wiki already has an edit history. Maybe you can do it during
editing, but once it's done, we should delete the "old" page, as it's
just confusing otherwise. I already submitted the application with the
link to https://github.com/sympy/sympy/wiki/GSoC-Ideas, so whatever is
on that page will be "the ideas page".
Aaron Meurer
> To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/CAHVvXxRcet5vmz91g0%3D33JFms2H59zbdiM0wau9ejSuvqi1vNg%40mail.gmail.com.
I agree with all of this in principle, but my experience is that it's
hard to get people to help out with updating the ideas page. And even
if I wanted to do it all myself, I don't know enough about some of the
ideas (like the ones in the physics submodule) to say anything about
them. I will say for a lot of the ideas I do know about, they are
actually still relevant, the issue is just that they are very
difficult so no one has ever been able to do them. It is worth keeping
them around in case someone actually comes along who can actually
manage them.
If you want to take a shot at doing this, you have my blessing. I
don't think it's necessary to move the current page to an "old" page
since the wiki already has an edit history. Maybe you can do it during
editing, but once it's done, we should delete the "old" page, as it's
just confusing otherwise. I already submitted the application with the
link to https://github.com/sympy/sympy/wiki/GSoC-Ideas, so whatever is
on that page will be "the ideas page".
Aaron Meurer
Oscar Benjamin
Jan 24, 2024, 6:16:48 PM1/24/24
to sy...@googlegroups.com
I'm not suggesting that we ask people to update the current content of
the ideas page. I am suggesting that we remove *everything* and then
say that if anyone wants something to be listed then they need to add
an up to date description along with a commitment that they will
mentor the project this year.
It doesn't matter if the old ideas are still relevant. If there is not
a mentor right now who understands what the current state is and
precisely what needs to be done then the project description is not
suitable for advertising to candidates.
Likewise if the idea has been on that page for 10 years with no one
doing it then probably the description of the project is no good.
Someone should turn it into a clearer, more useful description of a
more realistic project. Otherwise it shouldn't be listed there.
I am happy to remove everything from the GSOC ideas page but
understand that that is what I mean: I will remove *everything* and
then only add back one or two projects that I am personally willing to
mentor. Anyone else who wants to commit to be a mentor can add up to
date project ideas but otherwise if it is not clear that there is an
available mentor then the project should not be listed.
We should not roll anything over from previous years automatically in
the way that is currently done either in terms of the list of mentors
or the list of projects.
--
Oscar
> To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/CAKgW%3D6Lrid0UFb_mXtseY_oALESjxLMCCiD8kkpevkNwpg6GDg%40mail.gmail.com.
the ideas page. I am suggesting that we remove *everything* and then
say that if anyone wants something to be listed then they need to add
an up to date description along with a commitment that they will
mentor the project this year.
It doesn't matter if the old ideas are still relevant. If there is not
a mentor right now who understands what the current state is and
precisely what needs to be done then the project description is not
suitable for advertising to candidates.
Likewise if the idea has been on that page for 10 years with no one
doing it then probably the description of the project is no good.
Someone should turn it into a clearer, more useful description of a
more realistic project. Otherwise it shouldn't be listed there.
I am happy to remove everything from the GSOC ideas page but
understand that that is what I mean: I will remove *everything* and
then only add back one or two projects that I am personally willing to
mentor. Anyone else who wants to commit to be a mentor can add up to
date project ideas but otherwise if it is not clear that there is an
available mentor then the project should not be listed.
We should not roll anything over from previous years automatically in
the way that is currently done either in terms of the list of mentors
or the list of projects.
--
Oscar
Aaron Meurer
Jan 24, 2024, 6:32:19 PM1/24/24
to sy...@googlegroups.com
A lot of the advanced ideas are important enough that if we found
someone willing and actually able to do them, then we should make an
effort to actually find someone who could mentor them. Particularly
the ideas relating to the polys like CAD, better Groebner bases, etc.
Just glancing at the page, I would be willing to mentor most of the
non-physics projects, and would probably re-add them if you did this.
The main exception is the SymEngine projects. Let's see if any of the
existing SymEngine mentors are willing to mentor on those projects.
Otherwise let's remove them. We should also probably remove the
LFortran projects as I think Ondřej generally mentors projects for
LFortran from another mentoring org.
I went ahead and deleted the SymPy Live and SymPy Gamma ideas. SymPy
Live is already redone with pyodide, SymPy Gamma still exists but
ought to be shut down, and interactive examples is something that
other projects are working on, and we can adopt once they have
something usable.
Maybe the "solvers" idea should be rewritten, but it's definitely an
important thing that can be worked on.
TBH, I think you could probably come up with ideas for most submodules
in SymPy, and probably find someone to mentor most of them. Quite a
few wouldn't really be that important, which is maybe why they aren't
talked about.
Of course, it all boils down to finding a good student for them as
well. I honestly think that's way more important than the actual ideas
anyway. Almost all the "low hanging fruit" has been completed over the
years, and we mostly only have the harder projects left. Should we try
to find more (relatively) easy project ideas to add to the list? Or
should we just raise our bar of who we accept and accept fewer people?
I personally don't really like the low hanging fruit ideas because
they generate a lot of low quality applications and noise, but at the
same time, we have definitely had some good people in the past who
have started with such projects.
Aaron Meurer
On Wed, Jan 24, 2024 at 4:16 PM Oscar Benjamin
> To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/CAHVvXxQzW8dqW7gP5pF2gr4qkU2cGzeBfZqwzNefrXyKWEKqhA%40mail.gmail.com.
someone willing and actually able to do them, then we should make an
effort to actually find someone who could mentor them. Particularly
the ideas relating to the polys like CAD, better Groebner bases, etc.
Just glancing at the page, I would be willing to mentor most of the
non-physics projects, and would probably re-add them if you did this.
The main exception is the SymEngine projects. Let's see if any of the
existing SymEngine mentors are willing to mentor on those projects.
Otherwise let's remove them. We should also probably remove the
LFortran projects as I think Ondřej generally mentors projects for
LFortran from another mentoring org.
I went ahead and deleted the SymPy Live and SymPy Gamma ideas. SymPy
Live is already redone with pyodide, SymPy Gamma still exists but
ought to be shut down, and interactive examples is something that
other projects are working on, and we can adopt once they have
something usable.
Maybe the "solvers" idea should be rewritten, but it's definitely an
important thing that can be worked on.
TBH, I think you could probably come up with ideas for most submodules
in SymPy, and probably find someone to mentor most of them. Quite a
few wouldn't really be that important, which is maybe why they aren't
talked about.
Of course, it all boils down to finding a good student for them as
well. I honestly think that's way more important than the actual ideas
anyway. Almost all the "low hanging fruit" has been completed over the
years, and we mostly only have the harder projects left. Should we try
to find more (relatively) easy project ideas to add to the list? Or
should we just raise our bar of who we accept and accept fewer people?
I personally don't really like the low hanging fruit ideas because
they generate a lot of low quality applications and noise, but at the
same time, we have definitely had some good people in the past who
have started with such projects.
Aaron Meurer
On Wed, Jan 24, 2024 at 4:16 PM Oscar Benjamin
Oscar Benjamin
Jan 24, 2024, 7:01:59 PM1/24/24
to sy...@googlegroups.com
On Wed, 24 Jan 2024 at 23:32, Aaron Meurer <asme...@gmail.com> wrote:
>
> A lot of the advanced ideas are important enough that if we found
> someone willing and actually able to do them, then we should make an
> effort to actually find someone who could mentor them. Particularly
> the ideas relating to the polys like CAD, better Groebner bases, etc.
These are good examples of badly described projects. The project goals
>
> A lot of the advanced ideas are important enough that if we found
> someone willing and actually able to do them, then we should make an
> effort to actually find someone who could mentor them. Particularly
> the ideas relating to the polys like CAD, better Groebner bases, etc.
should be much smaller and the description should be much more
concrete about what is already implemented and what more can
reasonably be done in the timeframe of GSOC.
"Cylindrical algebraic decomposition" is far too big for a GSOC
project. The project description literally says:
Cylindrical algebraic decomposition
- Idea
-- Implement the Cylindrical algebraic decomposition algorithm
-- Use CAD to do quantifier elimination
-- Provide an interface for solving systems of polynomial inequalities
We can't possibly expect a student in a 175hr project to do anything
useful with this description. Coming fresh to the codebase and the
general topic of CAD it would take 175hrs just to figure out what
alternatives and prerequisites are already implemented in SymPy and
what the first steps towards implementing CAD for SymPy should be. We
can't expect a GSOC applicant to figure all this out for themselves as
part of the project (or even before applying!).
What should be listed as a project description is a much smaller
subproject that helps towards the broader goal of implementing and
making use of CAD. That subproject should be described by someone who
has a clear idea of the steps needed in the bigger picture and is
willing to mentor. Otherwise a project idea that says "Implement the
Cylindrical algebraic decomposition algorithm" is worse than useless:
we waste people's time by even suggesting that it is a reasonable
project.
--
Oscar
Sangyub Lee
Jan 27, 2024, 1:43:51 PM1/27/24
to sympy
> I personally don't really like the low hanging fruit ideas because
they generate a lot of low quality applications and noise, but at the
same time, we have definitely had some good people in the past who
have started with such projects.
I only see 'noise' that GSOC seasons makes the mailing list flooded with introductions, mentor requests.they generate a lot of low quality applications and noise, but at the
same time, we have definitely had some good people in the past who
have started with such projects.
I don't think that it is because we have some 'easy' ideas or not.
Is it really requested from GSOC to request students to share all the format, like instruction in the mailing list?
Or is it our own rule making this?
Is it really requested from GSOC to request students to share all the format, like instruction in the mailing list?
Or is it our own rule making this?
Anderson Bhat
Feb 10, 2024, 8:37:10 AM2/10/24
to sympy
Hello All,
I've updated the ideas for the topic "Improving Series Expansions & Limit Computations" (which is the project I worked on during GSoC 2022) for GSoC 2024. I think I've explained it enough for anyone to understand and frame a proposal. If anyone has any doubts they could reach out to me at anutosh...@gmail.com
Regards,
Anutosh Bhat
IIT Madras (2019 - 2024)
I've updated the ideas for the topic "Improving Series Expansions & Limit Computations" (which is the project I worked on during GSoC 2022) for GSoC 2024. I think I've explained it enough for anyone to understand and frame a proposal. If anyone has any doubts they could reach out to me at anutosh...@gmail.com
Regards,
Anutosh Bhat
IIT Madras (2019 - 2024)
Anurag Surendra Bhat (B20CS097)
Feb 11, 2024, 4:14:35 AM2/11/24
to sympy
Hello All,
Having come across multiple students on personal mail and SymPy's mailing list interested in contributing and eventually proposing for the control module project, I have opened up an issue compiling all the ideas which I would like to be worked upon for GSoC'24. It contains all the latest ideas, linked with their issues and relevant information.
I have added my name to potential mentors for GSoC'24 and I'm interested in mentoring projects improving the Control, Series and Stats modules. If anyone has doubts, they can reach out to me at anuragbh...@gmail.com
Regards,
Anurag Bhat.
Aaron Meurer
Feb 12, 2024, 4:55:46 PM2/12/24
to sympy
Another thing I forgot to mention: new this year is small projects,
which are 90 hour projects, in addition to the 175 and 350 hour
projects. Right now, none of the ideas specify 90 hours, but we should
probably add some smaller projects to the list.
Aaron Meurer
which are 90 hour projects, in addition to the 175 and 350 hour
projects. Right now, none of the ideas specify 90 hours, but we should
probably add some smaller projects to the list.
Aaron Meurer
Samith Kavishke
Mar 3, 2024, 1:31:48 AM3/3/24
to sympy
Is this concluded, does that mean all the current projects that can be seen in the Ideas page are available for students to apply?
(specifically regarding the Wolfram Interpreter)
(specifically regarding the Wolfram Interpreter)
Aaron Meurer
Mar 3, 2024, 4:01:26 AM3/3/24
to sy...@googlegroups.com, Francesco Bonazzi
I've Cced Francesco Bonazzi about the Wolfram Interpreter idea. It
relates to the RUBI project, but I don't know whether it is something
that is still needed for it. Personally I would prefer to avoid
creating a Wolfram Language interpreter if possible.
Aaron Meurer
relates to the RUBI project, but I don't know whether it is something
that is still needed for it. Personally I would prefer to avoid
creating a Wolfram Language interpreter if possible.
Aaron Meurer
> --
> You received this message because you are subscribed to the Google Groups "sympy" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to sympy+un...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/cf6c7cf3-82f4-435d-abd0-5b2d5e81b5dcn%40googlegroups.com.
> You received this message because you are subscribed to the Google Groups "sympy" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to sympy+un...@googlegroups.com.
Samith Kavishke
Mar 3, 2024, 4:08:43 AM3/3/24
to sympy
If that so, what are the ideas that we can really contribute? I hope "Enhancing the flexibility of MatchPy" is still needed.
Aaron Meurer
Mar 4, 2024, 3:10:27 AM3/4/24
to sy...@googlegroups.com
Yes, that idea is new this year.
Aaron Meurer
> To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/c95486ee-0523-4101-be26-b3d4a5d059b1n%40googlegroups.com.
Aaron Meurer
Reply all
Reply to author
Forward
0 new messages