This is the mail archive of the ecos-discuss@sources.redhat.com mailing list for the eCos project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Red Hat backs away from eCos?


On Thu, 20 Jun 2002, Kjell Svensson wrote:

<snip>

> 
> "Jim" (jyl087@netscape.net) suggested that chip manufacturers should 
> form a coalition group sponsoring further eCos development. While I 
> believe this is a beautiful idea, my experience with silicon 
> manufacturers unfortunately makes me pessimistic to whether this would 
> be realistic.
> 
> Instead, I'd suggest that we who are today part of the eCos user 
> community join together, founding and sponsoring an eCos startup company 
> who could safely bring a truly open-source eCos further!
> If we move on fast, hopefully we could maybe even attract some of the 
> original eCos developers to this eCos startup before they will have to 
> switch over to other businesses
> While I doubt such a company would be able to produce revenue levels 
> interesting enough to a VC, I'm positive that it could safely provide a 
> sound financial ground for a smaller number of developers continuing the 
> eCos development in a truly open-source spirit. And with the help from 
> contributions from a more dedicated "sponsorship community", I believe 
> the speed of furter eCos development could even be improved.

>From my perspective, this approach will fail for the same reasons a Redhat 
sponsored eCos has failed. As a new comer to eCos I can cite the reasons 
why I find it attractive:

1) Cheaper than any other Embedded RTOS. 
2) There is a mature tool-suite.
3) It's relatively easy to evaluate

I suppose many developers will cite these reasons. Now Redhat's eCos 
business model was to sell services such as porting eCos to custom 
hardware, writing drivers, or more thorough technical support. eCos, from 
my perspective, has matured to the point that porting to new hardware and 
writing drivers is a relatively simple matter of looking around for a 
project that is similar enough to yours and begin using it as a template.

For the technical side, it's clear from the traffic that I've seen on this 
list (and, I might add, have generated) that many eCos users attempt to 
solve the technical issues by wading through the details and asking very 
narrow, focused questions. In other words, it's not apparent to me from 
the messages I'm seeing on this list that people are actively using GNUpro 
and the support that comes with it.

In my personal case, once I get my eCos-based application up and running 
there is a chance others at my company will have to maintain it. At that 
point, I'd like to make that as easy as possible for them. In this 
instance, I might consider purchasing a professionally package tool. But 
then again, I might not...

In short, I can my job done without any extra cost.


> I suggest anyone interested in helping sponsoring an eCos startup 
> (indications from users about being willing to buy servicess from such a 
> startup would probably be the most valuabe item!) tell this to this 
> list, so that we can see it this could be a possible route to take.

As much as I would really *like* to agree with you, I don't see this 
happening. Part of eCos' fundamental appeal is that one doesn't have to 
purchase any services!

I hate to see Redhat withdrawing financial support for eCos, but I can see 
why. eCos clearly has achieved a critical mass in the embedded OS market. 
In my experience with Open Source software, that's adequate reason 
enough to keep "the project" alive. As users of eCos I think there are 
two main avenues left: 1) pressure the hardware manufacturers to provide 
an eCos Port, 2) contribute back to the project (with software patches).

Reason (1) works. At the early stages of my project I'm still evaluating
which processor that I'm going to use (although I'm almost certain it'll
be an Atmel ARM). The volumes for the project are significant - 100k's per
year. When I cited to an ARM competitor that one of the reasons I'm
looking at ARM is *because* of eCos (and the GCC tool chain), they made a 
(big) note of that and relayed that info "to the factory".

If these companies are lurking on the list by any chance, then my advice 
to them is that you ought to hire the eCos developers to port to your 
hardware/evaluation boards. The value of "free" software in this case is 
the hardware revenue that's generated. 

Reason (2) works as well. I know that I'll contribute back and I've seen
others doing so as well. This self-sustaining inertia has worked on many
Open Sourced projects - even the little ones that I happen to run. Unless
another embedded open source OS comes up that's significantly better than
eCos, I expect eCos to be active and around for a long time...


Scott


-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]