Hey Thiago,
I fully agree with your sentiment. I think the main problem is that most of
the current maintainers don't have the time to properly check the pull
requests. I am a bit careful with merging things because I did merge PRs in
the past that lead to things not working anymore and so I had to spend more
time fixing things up later.
I agree with your suggestions and would add a third one, namely to add more
maintainers:
1. Set up proper tests (a few datasets for English and other languages)
2. Create a protocol ("these things need to be checked before a PR is
merged")
3. Add more people as maintainers with the right to close PRs (I am
thinking: you, Sandro, David)
Jo
On Wed, Jan 21, 2015 at 2:31 PM, Thiago Galery <[email protected]> wrote:
> Hi there, first of all apologies for cross-posting. My name is Thiago and
> I have been using spotlight for a couple of years now. Although I think it
> is a great product and has a great community of people involved, I'd like
> to discuss some issues related to the maintainability of its codebase.
>
> I think that there s a lot of amazing developers out there that spend
> their evenings contributing with PRs to spotlight, but once they submit
> their work and maybe a discussion is going on, the PRs are not merged and
> it is not clear why that branch hasn't been merged. Is it because it lacks
> tests ? Or is it because it's a bad idea ? Or is it because people somehow
> get a bit scared of merging because they want to preserve the spotlight
> branch as clear/stable as possible so experimental work is unwelcome.
>
> To be honest, I don't know what's the best way to approach this is. But to
> illustrate this point further, here are a couple of PRs that just fix some
> functionality/make the code cleaner/more maintainable
>
> https://github.com/dbpedia-spotlight/dbpedia-spotlight/pull/289
>
> https://github.com/dbpedia-spotlight/dbpedia-spotlight/pull/312
>
> Another PR that is really cool is Dirk's Scala 2.10 + Factorie NER
>
> https://github.com/dbpedia-spotlight/dbpedia-spotlight/pull/306
>
> but in is this case, one could argue that 2 PRs should be submitted, one
> for the scala migration, another for the Factorie introduction.
>
> In any case, I think the following things should be in place, if the
> contributions from the community are welcome:
>
> 1. Some form of release metodology (development branch which PRs should be
> submitted against, release branch that goes through a proper load testing)
>
> 2. Automatic evaluation against a variety of datasets (these seem really
> useful https://github.com/diegoceccarelli/dexter-eval
> https://github.com/marcocor/bat-framework and
> http://acube.di.unipi.it/bat-framework/)
>
> 3. More clarity on why a PR is not merged yet/.
>
> If things don't improve, my worry is that people will either stop
> contributing or else focus on their own forks (which creates a problem of
> syncing master branches and what not).
>
> Comments thoughts and suggestions are welcome and appreciated.
> All the best,
> Thiago Galery
>
>
> ------------------------------------------------------------------------------
> New Year. New Location. New Benefits. New Data Center in Ashburn, VA.
> GigeNET is offering a free month of service with a new server in Ashburn.
> Choose from 2 high performing configs, both with 100TB of bandwidth.
> Higher redundancy.Lower latency.Increased capacity.Completely compliant.
> http://p.sf.net/sfu/gigenet
> _______________________________________________
> Dbp-spotlight-developers mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/dbp-spotlight-developers
>
>
------------------------------------------------------------------------------
New Year. New Location. New Benefits. New Data Center in Ashburn, VA.
GigeNET is offering a free month of service with a new server in Ashburn.
Choose from 2 high performing configs, both with 100TB of bandwidth.
Higher redundancy.Lower latency.Increased capacity.Completely compliant.
http://p.sf.net/sfu/gigenet
_______________________________________________
Dbp-spotlight-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dbp-spotlight-users
