Re: [Kiwix-developer] Kiwix and Evopedia (another offline wikipedia reader)

[Emmanuel E. <ke...@ki...>]

Hi

On 19.04.2015 10:38, Mossroy wrote:
> I'm another contributor of Evopedia (I mainly developed the Firefox OS
> port).
> I share what Christian said here (we discussed it together before
> sending the email).
> See my few comments in the text below.
>
> Le 18/04/2015 13:37, Emmanuel Engelhart a écrit :
>> The Kiwix team, of course, knows about the Evopedia project which is a
>> really interesting and good software from many point-of-views (from my
>> personal point-of-view, your work around FFOS or Qt is in particular of
>> interest).
>>
>> It's sad to read that you face a (hopefully temporary) lack of
>> resources, but this is a problem we know pretty well too.
>
> My personal opinion is that we should try to share
> ideas/resources/code/maintenance in order to avoid spliting them between
> 2 different projects that have more or less the same goals.
> We both use Free Software licenses, which should help sharing code.
>
> I can not say if this "merge" might work or even be worth it.
> But I think it's worth trying.

Yes, I agree. Kiwix, as a project, already faced a similar situation in 
the past with the project Moulinwiki. To my opinion, the start&important 
point is tech. collaboration. For the rest, this is like a Mayonnaise, 
it works or it doesn't...

>> FFOS is from our point of view probably the most promising development
>> in the mobile sector since a long time. We are really interesting by
>> having a Kiwix version for FFOS. That said we don't have invested a lot
>> of time until now on it. We were in touch with someone who had done a
>> ZIM reader in javascript and this was working. The problem was the
>> performance... but we have already in mind a few improvements which
>> should be able to improve the situation.
> Whoa, that's pretty interesting. Could you please point us to this
> javascript implementation?

This was the first question of Christian too! The answer is that the 
javascript was generated from the C/C++ code directly and the code was 
not put online (but we are in touch with the developer).

> There also are performance issues in the javascript implementation of
> Evopedia. The main concern is to have a fast javascript decompresser
> that can work on small chunks.

Yes. Challenging and at the same time really interesting topic.

>> To conclude: FFOS is a really strategic move we have to do, but for now
>> and in 2015, there is no real plan to work on that.
> I also think that Firefox OS is a great platform.
> IMHO, it makes sense to have an offline wikipedia reader on cheap
> smartphones in emerging countries (where people will more likely be
> using a smartphone than a standard computer). And it's precisely the
> initial market target of Firefox OS.

100% agree.

> Plus an HTML5/javascript application might be run on virtually any
> platform (PC, Mac, iOS, Android, Windows Phone, Blackberry etc), through
> Apache Cordova when necessary.

Although this idea looks brilliant from the theoretical point-of-view, 
the reality of the implementation is IMO more complicated. For example 
the Wikimedia Foundation has abandoned that way (Cordova as mobile 
cross-platform solution).

After thinking pretty much about this, our approach at Kiwix, for now, 
is to have as much as possible code in C++ and then use native SDK for 
the UI. Of course this does not work for FFOS.

The only signal I see that we might change this approach in the future 
is he success of an application like PopcornTime. Popcorn time seems to 
have achieved to build a successful cross-platform (for both 
desktop&mbile) app based on javascript (but I guess part of the core is 
still binary).

>> Like you probably know, Kiwix is mostly a volunteer driven project, so
>> anybody is more than welcome, in particular for this kind of features
>> are really important ones.
> I'll contribute as much as I can.
> In particular, I might be of some help about Firefox OS (I'm a bit
> involved in this community).
> But, in the short term, I unfortunately won't have a lot of time to code.

Yes, it looks like FFOS is a good entry point for collaboration.

Thx for your nice email.
Emmanuel

-- 
Kiwix - Wikipedia Offline & more
* Web: http://www.kiwix.org
* Twitter: https://twitter.com/KiwixOffline
* more: http://www.kiwix.org/wiki/Communication