You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@spamassassin.apache.org by Sarang Shrivastava <sa...@gmail.com> on 2019/04/05 16:35:16 UTC
Re: Gsoc 2019 SpamAssassin Proposal
Hi Kevin,
I went through the proposal. It looks crisp. I had a few things which were
still unclear to me and was not sure if we should mention these things in
proposal or not ?
1) The statistical ML models can be implemented in python as it has
enormous amount of library support for the same. We can then use this
trained model to pass in scores for a new email via a plugin in perl. Do
you think this is feasible ? Do you have bay alternative approach in mind
by which we can do this ?
2) Lets say that we have mentioned some models in our proposal and they
don't perform that well in the actual evaluation phase. In that case are we
flexible enough to try new models on the fly which weren't even mentioned
in the proposal ?
Cheers,
Sarang
On Sun, Mar 31, 2019, 6:44 AM Kevin A. McGrail <km...@apache.org> wrote:
> I could nitpick things here and there but to me, I really liked the
> schedule.
>
> I might recommend you add something like, if accepted, the project has
> agreed to support me and identified Kevin A. McGrail as a mentor.
>
> Regards,
> KAM
>
> On 3/30/2019 12:00 PM, Shreyansh Shrivastava. wrote:
>
> Hello everyone,
> This is Shreyansh Shrivastava, Computer Science and Engineering student at
> NITK India.
> I am interested in working on the below-mentioned project during GSoC
> 2019.
>
> SpamAssassin Statistical classifier
> <https://issues.apache.org/jira/browse/COMDEV-268?jql=text%20~%20%22GSOC%202019%20SpamAssassin%22>
>
> I have finalized the first draft of my proposal. Any further
> suggestions/changes in the proposal are highly appreciated. Please add
> comments to Google doc wherever you feel appropriate.
>
> Gsoc 2019 Proposal
> <https://docs.google.com/document/d/16rDbjhtUmJKDgmLL_76C0vVm6SYRPjhnc5w5R1cen04/edit?usp=sharing>
>
> Regards,
> Shreyansh Shrivastava
>
>
> --
> Kevin A. McGrail
> Member, Apache Software Foundation
> Chair Emeritus Apache SpamAssassin Projecthttps://www.linkedin.com/in/kmcgrail - 703.798.0171
>
>
Re: Gsoc 2019 SpamAssassin Proposal
Posted by "Shreyansh Shrivastava." <sh...@nitk.edu.in>.
Hey Kevin,
I have tried to include all the suggestions. Also, the testing phase is
added in the proposal.
Since only two days are left for the proposal submission deadline, I would
really appreciate any further pointers from you and the community to make
the proposal better.
Gsoc 2019 Proposal
<https://docs.google.com/document/d/16rDbjhtUmJKDgmLL_76C0vVm6SYRPjhnc5w5R1cen04/edit?usp=sharing>
Regards,
Shreyansh Shrivastava
On Sat, Apr 6, 2019 at 9:34 PM Kevin A. McGrail <km...@apache.org> wrote:
> On 4/6/2019 10:33 AM, Shreyansh Shrivastava. wrote:
> >
> > I realised that my proposal doesn't have any dedicated content for
> > testing of the plugin.
> >
> > Firstly, should I include it in my proposal ?
> >
> > Secondly, what should be the flow of the testing section if at all we
> > are including it ?
>
> I would say yes, it doesn't hurt to show you are paying attention to
> the full cycle of software development. For me, I usually try and
> write the tests first and then the code. You could likely just say you
> will be using test-driven development.
>
> --
> Kevin A. McGrail
> Member, Apache Software Foundation
> Chair Emeritus Apache SpamAssassin Project
> https://www.linkedin.com/in/kmcgrail - 703.798.0171
>
>
Re: Gsoc 2019 SpamAssassin Proposal
Posted by "Kevin A. McGrail" <km...@apache.org>.
On 4/5/2019 12:35 PM, Sarang Shrivastava wrote:
> Hi Kevin,
>
> I went through the proposal. It looks crisp. I had a few things which
> were still unclear to me and was not sure if we should mention these
> things in proposal or not ?
>
> 1) The statistical ML models can be implemented in python as it has
> enormous amount of library support for the same. We can then use this
> trained model to pass in scores for a new email via a plugin in perl.
> Do you think this is feasible ? Do you have bay alternative approach
> in mind by which we can do this ?
Sure, using python is feasible though I won't be able to help much with
python. You'll be teaching me.
Can you restate your bay alternative question?
> 2) Lets say that we have mentioned some models in our proposal and
> they don't perform that well in the actual evaluation phase. In that
> case are we flexible enough to try new models on the fly which weren't
> even mentioned in the proposal ?
Sure but I would say to set your goals on what's in the proposal. Time
management is the reason for most failures in the program.
Regards,
KAM
>
> Cheers,
> Sarang
>
> On Sun, Mar 31, 2019, 6:44 AM Kevin A. McGrail <kmcgrail@apache.org
> <ma...@apache.org>> wrote:
>
> I could nitpick things here and there but to me, I really liked
> the schedule.
>
> I might recommend you add something like, if accepted, the project
> has agreed to support me and identified Kevin A. McGrail as a mentor.
>
> Regards,
> KAM
>
> On 3/30/2019 12:00 PM, Shreyansh Shrivastava. wrote:
>> Hello everyone,
>> This is Shreyansh Shrivastava, Computer Science and Engineering
>> student at NITK India.
>> I am interested in working on the below-mentioned project
>> during GSoC 2019.
>>
>> SpamAssassin Statistical classifier
>> <https://issues.apache.org/jira/browse/COMDEV-268?jql=text%20~%20%22GSOC%202019%20SpamAssassin%22>
>>
>> I have finalized the first draft of my proposal. Any further
>> suggestions/changes in the proposal are highly appreciated.
>> Please add comments to Google doc wherever you feel appropriate.
>>
>> Gsoc 2019 Proposal
>> <https://docs.google.com/document/d/16rDbjhtUmJKDgmLL_76C0vVm6SYRPjhnc5w5R1cen04/edit?usp=sharing>
>>
>> Regards,
>> Shreyansh Shrivastava
>
>
> --
> Kevin A. McGrail
> Member, Apache Software Foundation
> Chair Emeritus Apache SpamAssassin Project
> https://www.linkedin.com/in/kmcgrail - 703.798.0171
>
--
Kevin A. McGrail
Member, Apache Software Foundation
Chair Emeritus Apache SpamAssassin Project
https://www.linkedin.com/in/kmcgrail - 703.798.0171