About the future of the freesmartphone.org middleware

[Simon Busch]

Hey everybody,

as a lot of you may have noticed we did two releases in the past months 
of the FSO stack. Both were related to bring stability and consistence 
to the stack. Now I want to talk with you about the future of the stack. 
In the past we were only concentrating on getting new hardware supported 
and lost our real focus on creating a middleware suitable for 
embedded/specific-purpose devices. This is where I want to go back to 
and get into development again.

In the last weeks I looked over several parts we have in our stack and 
tried to find out where we can improve and get into development of new 
features again. A lot of you have stability in mind as you want to use 
something with FSO on your daily phone. Thats the second peace which 
should be part of the core development focus of the Freesmartphone.org 
middleware.

Getting new features is fast said but I have several things on my list 
where I want to improve FSO in the next weeks and months. Everything is 
focused on the core stack which is formed by our framework libraries and 
the three daemons fsodeviced, fsogsmd and fsousaged. We have other 
daemons like fsotdld as well but that will be not on my focus. If 
someone wants to step up with further development of these just go on 
and get in contact. But please don't get me wrong: I will support all 
other daemons like fsoaudiod and fsotdld in the next releases too but 
just not doing any development related work for them.

For fsogsmd there are the following things on my list:

1. Get the last peaces of not implemented things in like conference or 
emergency calls
2. Several API cleanups
3. Get several bugs fixed
4. Do integration testing with a remote controlled phonesim so we can 
simulate incoming calls etc. This will also included integration testing 
with a remote controlled fsogsmd on another device
5. Multi device support: While working in HFP HF support in fsogsmd I 
discovered that things would be easier if we can control more than one 
modem with the same daemon at the same time. Think about phone with 
support for more than one SIM card. Work has already started for this in 
the morphis/multi-device branch of the cornucopia repository.
6. Cleanup of the modem status handling: right now the modem status and 
SIM/network status are too much tight together. We have cleanly separate 
them.
7. Internally we don't separate a modem from a AT based modem; that 
needs to be fixed
8. A lock-down mechanism to keep anyone out when doing a firmware 
upgrade. When doing a firmware upgrade of a modem we have the problem 
that nobody should access the modem while this is in progress. The idea 
is now to implement a dbus API to lock the modem by requesting a lock 
and only the requesting program can unlock the modem again. While the 
modem is locked nobody else can access the modem via fsogsmd.

fsousaged:
- nothing right now

fsodeviced:
- nothing right now

lib*:
- I am thinking about grouping all libraries together so just have one 
single framework library and don't need to maintain several small 
libraries which increases the complexity of release testing, ABI 
refinements, etc. Just one libfsoframework; this is only a thought in my 
head right and nothing concrete.

That are some of my toughs were I want to go with FSO in the next 
months. So no focus on concrete devices but getting the stack itself 
forward to be ready for every kind of a device.

I would be really happy to hear what other people are thinking about the 
idea behind FSO since it was started back in 2008. What are your missing 
features? What do you like and what not?

regards,
Simon

-- 
Simon Busch - http://mm.gravedo.de/blog/