Reverse or Creation?

Dr. H. Nikolaus Schaller hns at
Fri Mar 14 19:42:49 UTC 2014

Am 14.03.2014 um 20:14 schrieb Paul Kocialkowski:

>> The free software and open source software works because software is less
>> prone to obsolescence than hardware which allows armies of dev to work in their
>> precious spare time on promising projects for years - even decades before
>> they begin to be relevant (at which point financing appears... ironically).

The only problem is that we (users) allow manufacturers to sell us a new device
every 6 months. Because we run behind the latest and greatest features.

>> You just can't do that with hardware. You must spend 150% of your time
>> to quickly win the race against obsolescence and huge proprietary industrial
>> producers and quasi monopoly. It's just not feasible. Maybe if research
>> centers decided (or were ruled to) open source, creative common, open
>> science or free software the results of their public funded results, I
>> could consider it.
> I totally agree. As a software developer working on embedded devices
> that become obsolete in a 6-month time frame (well, that's according to
> the industry, people make their devices last longer), I find it really
> depressing how the software we write for a particular device can become
> nearly useless when the device becomes obsolete (even though it's good
> to have it as a reference for other devices, but that's a
> chicken-and-egg issue).

Yes, that is what we want to make completely different with the OpenPhoneux
initiative: have devices that are not obsolete after 6 months because they
are abandoned by the manufacturer.

There is still support for the GTA02 or GTA04 after years. And since a lot of
information is really open it can be supported.

And what is rearely understood: yes such a GTA0x is more expensive
than a modern device. But if you keep it for 4 or 5 years and get support
instead of buying 3 less expensive unsupported devices over the same
time, total cost of ownership is lower.

Even if you subtract the effect of not being able to show around the latest
and greatest design and features.

> I suppose one of the biggest problems with hardware is the need of
> producing a new device to implement a new feature. I don't see any easy
> way to fix this. Even if we decided to stick with a device for several
> years, knowing its functionalities are not going to evolve, we'd still
> be stuck with software being more and more demanding (in accordance to
> new devices' abilities).

That is not a necessity and it *can* be done differently. I have an example of
(old) Apple (when they were still interested in their customers):

In 1993 I bought a 68040 based PowerBook with Mac OS 7.something and
Framemaker 4 for doing professional scientific text processing.

10 (!) years later I was able to install and run the same binary application on
a PowerPC and run it in an emulator on OS X 10.2 or 3 (don't remember exactly).

And believe it or not - it was faster than ever and even did fix a bug with font rendering.

So if software and API developers want to keep the APIs stable and running, it is
no problem to keep software long living.

But I must admit that I don't see how *we* can do that for Replicant, because
it is driven by Android which is driven by Google and they are not that much
interested in keeping APIs stable for a long time (like Apple isn't nowadays).

> For instance, take the latest Android versions:
> it's harder and harder to make them run on "old" Android devices such as
> the Nexus S, not to mention the HTC Dream. That's not only the case with
> Android: look at GNOME-Shell abandoning its EGL-less fallback (AFAIK the
> fallback gnome-shell is too slow to be usable). However, some software
> still exists (enlightenment), but adapting it and a whole GNU/Linux
> system for mobile devices (i.e. SHR) is more work that porting Replicant
> to a new version on a new device.
>> My 2 cents
>> Sebastien
>>>> Than, let me ask - wouldn't it be better to design a NEW system
>>>> from
>>>> scratch instead of spending years on reverse-engineering of what
>>>> manufacturers are selling?
>>> The real problem here is that we won't be able to have community
>>> chips
>>> manufacturing in the near future. The best we can do is to gather
>>> chips
>>> from free software-friendly manufacturers and assemble devices that
>>> way.
>>> This is already done with e.g. the GTA04/neo900.
>>> Devices with good design also exist (e.g. Allwinner tablets/single
>>> board
>>> computers) and are about as good as what you can get with the GTA04.
>>>> Personally i chose not to spend time on reverse stuff. Concentrated
>>>> all my power on research for new design.
>>> From a developer point of view, starting a new platform from scratch
>>> is
>>> not necessarily *less* work, it's just different work. Reverse
>>> engineering stuff is currently doable (on selected platforms) and
>>> well,
>>> if we're already going that way, it's because it's the easiest path
>>> to
>>> freedom currently.
>>> --
>>> Paul Kocialkowski, Replicant developer
>>> Replicant is a fully free Android distribution
>>> Website:
>>> Redmine:
>>> _______________________________________________
>>> Replicant mailing list
>>> Replicant at
> -- 
> Paul Kocialkowski, Replicant developer
> Replicant is a fully free Android distribution
> Website:
> Redmine:
> _______________________________________________
> Replicant mailing list
> Replicant at

More information about the Replicant mailing list