[Opensync-devel] OpenSync: fragmentation is harmful

[Daniel Gollub]

On Tuesday, January 04, 2011 12:42:55 pm Patrick Ohly wrote:
> Several reasons:
>       * libsynthesis did not (and still doesn't) fit into the OpenSync
>         architecture because engine and data conversion are closely
>         related. Arguably this is necessary, because one needs
>         information from the other (capabilities, choosing data
>         formats). Refactoring libsynthesis so that it fits into the
>         OpenSync plugin concept of formatters, backends and engine would
>         have been a major effort, both on libsynthesis and also
>         OpenSync.

Right. OpenSync is currently still working towards that until OpenSync or your 
or someone else solution full filled the goals OpenSync is targeting.


>         As you know, we discussed it at the time, but despite
>         all the emails that we exchanged I still did not fully
>         understand the OpenSync design and thus did not pursue this
>         further.

I see. I'm sorry about that ... I invested a lot of time in supporting you as 
much information i was able to do in my limited time ...

Synchronization is complex as you now. But this might also happen to 
SyncEvoluation that you will not get more contributors because nobody fully 
understand - beside you - how SyncEvoluation is designed.

For OpenSync, i haven't done the initial design - and also managed to 
understand most of the implementation without contributing major 
implementation like Armin and many others did.

So i just can give you the advice to make really really really good 
documentation. E.g. just having providing backend-samples is not enough.

For OpenSync Documentation of the internals will be very likely the next thing 
I'm going to do, to enable other people to finish 0.40.


>       * State of OpenSync. We needed a fully working solution within a
>         few months for Moblin 1.0 beginning of 2009. Two years later
>         there still isn't a stable OpenSync release, so I think the
>         decision was justified. I also don't think that investing work
>         into OpenSync would have led to a different result (pure
>         speculation, of course - we'll never know).

You need to contribute to projects if you want to gain something particular 
out of it.

> 
> > Something I'm very interested and also prepared several changes for, and
> > added  additional complexity, to introduce libsynthesis in OpenSync.
> > E.g. by making OpenSync independent of xmlformat.
> 
> Unfortunately refactoring libsynthesis is still blocking that, and there
> is no-one working on it. It simply isn't important enough because
> libsynthesis works well for its current usage.

I could imagine that it would be much more valuable for the FOSS community to 
re factor libsynthesis and OpenSync to have a lot common code base.

While developing OpenSync for several years i got in touch with a lot of 
people developing PIM interfaces. And there is so much redundant code which 
could be common code base. libsynthesis is very mature and well maintained 
code and would be much more of use for the FOSS community if it would be 
refactored. Which might be not important for you, but for other projects.



Currently i don't see syncevolutaion is replacing OpenSync today. Maybe it 
will some when in the future. Maybe not. It could disappear like other Sync 
approaches due to various reasons. I'm not going to declare OpenSync as dead. 
I highly recommend that OpenSync (Plugin) Developers decide by their own if 
syncevoluation is a better alternative for them or not. It would be pretty 
ignorant to say anything different.

My personal goal is to support the FOSS community so they have some tool to 
have a rock-solid synchronization solution. I will not declare OpenSync as 
dead in favor of SyncEvoluation. Not only because I'm not in position to 
declare the project as dead. Because I'm not convinced that SyncEvoluation is 
an alternative to OpenSync to date. And also because i have the impression 
that there are still lot of people which still want contribute to OpenSync and 
those I want to support with my contribution.

I doubt OpenSync is a "honey pot". Maybe your project is for various reason 
just not attractive enough. Maybe renaming it, like suggested by Chris would 
help.


The only harmful thing for both project is - discussing that fragmentation is 
harmful ;)

Instead each of us could concentrate on working towards their implementation 
make it just work.


-- 
Daniel Gollub
Linux Consultant & Developer
Mail: gollub at b1-systems.de

B1 Systems GmbH
Osterfeldstraße 7 / 85088 Vohburg / http://www.b1-systems.de
GF: Ralph Dehner / Unternehmenssitz: Vohburg / AG: Ingolstadt,HRB 3537
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <https://lists.freedesktop.org/archives/syncevolution/attachments/20110104/b508ea86/attachment.sig>