Re: Timeline for GSoC 2017 project (chat history improvements)

Paulo Lieuthier <[email protected]> Sun, 28 May 2017 23:04:57 -0300
Newsgroups	gmane.comp.kde.devel.kopete
Message-ID	<[email protected]>
On 27-05-2017 06:12, Pali Rohár wrote:
> On Friday 26 May 2017 16:41:15 Paulo Lieuthier wrote:
> > Hi all,
> >
> > I'm really sorry it took this long. I've been really busy with my
> > undergraduate thesis, as I'm graduating this semester. The final
> > date for my thesis is June 12, very close GSoC's first evaluation.
> > This is going to be hard!
> >
> > Anyway, I have prepared a tentative timeline [1] that I think is
> > reasonable, but I really need everyone's opinions and expectations
> > about it. Is it sufficient? Please review it and comment on the
> > document.
> >
> > I will update the timeline during the project so it stays always up
> > to date.
> >
> > Thanks,
> > Paulo Lieuthier
> >
> > [1] https://goo.gl/vs7dwG
>
> Hi! I'm not quite sure if history2 is a good starting point. In whole
> timeline is missing one important part: create database/relation schema
> (or whatever would be used) and how messages would be stored/serialized
> into storage. This needs to be designed as a first part, because whole
> implementation depends on it. Also incorrect db schema would mean wrong
> performance and unusability. In timeline is also missing time for
> writing conversion from existing XML plugin to the new one.
>
> PS: In future please send timeline or other documents to mailing list
> (plain text is enough), not some external link which does not have to
> work without internet connection.
>

Hey, Pali. I thought history2 was the best option today. I just updated my
timetable aiming to convert the XML-based one, and start by preparing the schema
and searching for a good database solution.

Do you think converting the SQLite-based plugin into another backend is worth it
(that if SQLite is not the chosen database solution)? I can do it if there is
enough time.

The updated timetable:

May 29 – June 2
* Become familiar with Kopete’s current history plugins’ codebases.
* Discuss with ML the reasoning for deprecating previous plugins, and also
features that should be kept and new ones.
* Get the XML-based history plugin to build on top of kf5 branch.

June 5 – June 9
* Wrap my around of a good and sufficient schema for the conversations history.
* See how other apps manage messages history.
* Search for a database solution, both schema-friendly and text search-friendly
(SQLite, Firebird, Xapian)?.

June 19 – June 23
* Add a façade abstraction for storage backend in history plugin.
* Add initial implementation of new storage backend alternative.

June 26 – June 30
* Add a UI prototype for chat history. Only Jabber will be supported by now.

July 3 – July 7
* First Evaluation.
* Add full text search to history UI, with typo tolerance.

July 10 – July 14
* Support all account types (direct messages, groups, channels).

July 17 – July 21
* Add configuration UI and discuss sensible options and defaults
* Work on automatic history retrieval when reaching beginning of conversations.

July 24 – July 28
* Second Evaluation.
* Add unit tests for storage backends.
* Update XML-based storage backend to support new functionalities (needs
discussion).

July 31 – August 4
* Work on more unit testing (configuration UI, backend switching, text search).
* Test special case scenarios (corrupted database file, account deletion and
recreation etc).

August 7 – August 18
* Dogfooding and edge polishing.

August 21 – August 29
* Final Evaluation

Let me know if there is still something to improve in it.

Paulo
signature.asc (application/pgp-signature, 833 B)