Red Hat will give eCos Copyrights to the FSF!

An anonymous reader notes "Businesswire reports in this article that RedHat will assign its copyrights for the eCos embedded OS to the FSF. This is great news, considering that they have stopped developing it in 2002. Hopefully this will mean new life for the project."

Comments

New life for the project?

[SpanishInquisition]

You mean like HURD?

If development stopped in 2002...

[Fortunato_NC]

how come it looks like they added CPU architectures and features in 2003?

Re:If development stopped in 2002...

[mcspock]

Development by redhat stopped in 2002, when they did a round of layoffs. Basically the entire ecos dev group (which all came from the cygnus buyout) got dropped, and the majority of them went to form eCosCentric.

Redhat has continued to host the eCos project, just like they do for gcc and gdb, and the eCosCentric team has been writing updates as far as i know.

Stopped developing it in 2002?

[Saeed al-Sahaf]

This is great news, considering that they have stopped developing it in 2002. Hopefully this will mean new life for the project."

The web site indecates new development as recent as September of last year.

Re:Stopped developing it in 2002?

[jifl]

>The web site indecates new development as recent as
>September of last year.

Um, development has been ongoing, irrespective of Red Hat's loss of interest back at the start of 2002. There just hasn't been any big news since then. See the patch list for example.

The eCos maintainers (of which I'm one) have been pushing for a solution to the copyright issue for quite some time. It's good for everyone that Red Hat have donated eCos to the FSF.

Depends

[AKAImBatman]

Hopefully this will mean new life for the project.

I guess that kind of depends on whether anyone cares or not. Most people who might have used eCos for the commercial support aspect, are using the high powered and rock-solid QNX OS. And those who wanted free embedded OSes for home projects are already using Embedded Linux or *BSD. Even more difficult for eCos is that embedded Linux and *BSD distros are usually custom to the application. Why would anyone want the overhead of a prepackaged solution?

Perhaps eCos has its uses, but it's a very small niche.

Re:Depends

[AKAImBatman]

You painted an interesting picture, but left a big gap.

There are a lot of commercial companies actually using Linux, for whom QNX and eCOS are not worth the investment in light of recent kernel advances.


I'm sorry. I could have sworn that I already stated that the other end of the spectrum is "already using Embedded Linux and *BSD". Oh wait. I did.

You might save yourself some trouble if you read more carefully. :-)

Re:Depends

[mcspock]

My last company used eCos to build handheld and stereo component MP3 players, so that's my experience with it. In the handheld space it was great; small footprint (i built 40K ram/200K rom mp3+wma players with it), low overhead, minimal MMU requirements. The problem, with that space at least, is that the entire segment has shifted to faster processors with better MMUs, bigger hard drives, and generally larger requirements, which warrants using Linux. Even the eCos team was aware of this, as they started adding support for CPUs with memory protection and implementing more advanced OS features, basically scaling eCos up to...a trimmed down Linux.

It was very good and extremely competitive at the time though; i think the issue is just that this time has passed.

Re:Depends

[jifl]

[Uh oh! Advocacy war storm clouds gather]

I think you need to read up more on eCos! To call eCos "prepackaged" is about as far from the truth as you can get. The big C in eCos stands for configurable, and it is far more configurable and customizable to your application than any Linux or BSD will ever be, and certainly QNX.

eCos is for the deeply embedded market, and embedded Linux, even in 2.6 is so much bigger. eCos systems start from just a few KB (~10KB I think I remember), and scale up from there as you use more features - using configuration, just exactly the features you want, and with the semantics you want. You get the choice.

Add to that that eCos is completely open source, and royalty free with no upfront costs either (although you do have the option of commercial support if you do want it), and you'll understand why eCos is so popular.

eCos supports many more targets and architectures than QNX too.

Abandonware

[Rhubarb Crumble]

Is this the first time a software developer has expressly relinquished copyright for abandonware? Of course, eCos was never proprietary, so it's not quite the same...

Re:Abandonware

[TwistedSquare]

I remember reading (can't find mention on the site though) that Al Lowe, creator of the Leisure Suit Larry series, released the copyright on various old games (the ones owned by him rather than the publishers) for abandonware, since otherwise they would have died out... Confirmation would be good though.

Re:tax writeoff

[greenhide]

Probably not, unless they can show lost profit due to this maneuver. I once tried to donate a few websites to some organizations. After I'd developed them, I found out that I can't deduct one dollar of their value. Not one. Basically, the only thing you can easily take a deduction for is hard goods or cash.

hmm...

[Suppafly]

I wonder what license the FSF will put on the copyrights when they get them?

Go Red Hat!

[fader]

As much as everyone seems to hate Red Hat for being big and pushing for what they want, I have to say that I truly respect them. They've never been anything but fair, honest, and helpful to the OSS community. Sure, they sometimes make unpopular decisions, but they focus on their business and don't try to meddle in anyone else's. And they've certainly never pulled any stunt worthy of calling them the 'Microsoft of Linux' as gets thrown about from time to time. The worst they've ever done is ask that people redistributing their distro use a different name and artwork. I can't see how anyone could have a problem with that.

And as this shows, they often go above and beyond the mere requirements of the GPL. They've released a good amount of software under the GPL when they really didn't have to. They pay a lot of developers' salaries, too.

So I'd like to say thanks, Red Hat. I have nothing but good feelings toward you, and I hope you do better and better financially.

(Full disclosure: I don't work for Red Hat, don't own any of their stock, etc. I knew one guy who worked for them, but he was a tech support grunt there for a few months and I wasn't even in contact with him then. These opinions are my own.)

Dreamcast Linux

[Erwos]

The Linux-Dreamcast port apparently uses eCos to do some of the initial booting. So, while I wouldn't say I've seen it used practically, it was a nifty application of the OS.

-Erwos

Mixed Feelings about news like this

[dr_canak]

I'm a Redhat shareholder, not a lot of shares, and bought well after they bottomed out. I bought the shares for two reasons.

The first was because I thought if any company had a shot a taking a piece of Microsoft, it was Redhat. I use MS products everyday, probably always will, but I'm one who believes that MS got where they are with unfair market practices and ended up a little too big for their (and our) own good. By buying Redhat shares, I figured I was backing up my philosophy with my wallet.

The second reason is purely more pragmatic. Assuming Linux can take a stab at Microsoft, I believe Redhat is the most viable company to do it. In 30+ years, I'd like to think my decision to buy Redhat shares (when it was $12.00 a share) will be similar to people who bought Cisco, Oracle, MS, etc... back in the day. Sure the stock prices have wildly fluctuated, but look at the splits, and you realize just how much money there was to be made. So of course I would like to see that kind of return on this investment.

Which is why I end up conflicted when I see news like this. On the one hand, giving away a copyright is exactly the kind of collaboration you see with the Linux development model, and why it *may* in the end surpass MS in some, if not all, applications. But as a shareholder, giving away copyrights is hardly a way to grow a business. It took time, money, and effort to secure the copyright. Who knows if this news really effected shareprice, but with the release of this news, Redhat is down almost .50/share. So as a stockholder hoping to make money on my investment, I'm not too thrilled with this kind of news.

I suppose that's why you need to leave your emotional mind out of the market place, to avoid investing with your heart, and not your head ;-).

just my .02 (- .50)
jeff

Re:Where has eCos been used?

[ams001]

There are loads of commercial products and projects using eCos. See http://www.ecoscentric.com/ecos/examples.shtml

What has really changed?

[pavon]

I'm not certain what effect if any this has on the development of the software. To correct several misinformed posts: As the article mentions Red Hat stopped development on the project in 2002. The community continued which is why you see new releases after then. Second, the software was already open source - the licence has not changed. What has changed is that they given copyright over to FSF. The reason for this is that it is easier from a legal standpoint for the copyright of a project to be held by a single entity who can defend the entire project rather than each little peice being copyright of the respective authors. Since Redhat was no longer actively developing eCos, it made since for them to turn over the copyright to someone else. But unless people were resistant to contribute because RedHat still maintained copyright, I don't see how this will give the project new life. What may help more is having the fact that the project has a new maintainer (and the front-page slashdot article won't hurt either ;)

That's nice. Customers should require this.

[Animats]

Big customers should require that when a company drops a software product, it goes open source. This offers the option to keep the product alive if it's needed by the customer. Such terms are occasionally seen in the embedded world, but on a single-customer basis. A standard, well-accepted contract for software escrow and open-sourcing when the product is abandoned would be a useful thing to have.

Vendors go bankrupt, exit a field of business, or simply discontinue products all the time. Deals like this could help small vendors, providing long-term customer assurance.

Re:This is strange.

[The Lord of Chaos]

AFAIK eCos was always published under the GPL.

Assigning copyright to the FSF means that the FSF now owns the eCos codebase and they can do whatever they want with it including publishing it under the GPL.

Basically the point of this is so that if a developer wants to contribute to the eCos codebase they fill out a copyright assignment to the FSF instead of RedHat from now on.

Re:Wonderful news!

[OrangeTide]

Not really big news. eCos was GPL since March of 2002. The difference is that Redhat is giving the copyright over to FSF to watch over. Prior to March 2002 eCos was under the Red Hat eCos Public License. If you go back even further eCos was first released in October 1998 by Cygnus Solutions. And [as far as I know] was property of Cygnus Solutions until November 1999, which was when Red Hat aquired them.

If you are interested in developing with eCos the only book I know of is
Embedded Software Development with eCos

First chapter of the book...

1.1 Where It All Started--Cygnus Solutions

Michael Tiemann, David Henkel-Wallace, and John Gilmore founded Cygnus Solutions in 1989. The idea behind Cygnus Solutions was to provide high-quality support and development for open source software. It was initially unclear whether this business model would work out; however, by the end of the first year it was obvious from the value of the support and development contracts that the business was real. The workload was enormous for the five-person company (the three founders, a salesperson, and a part-time graduate student).

It was clear that the engineering support model worked; however, the costs to fulfill these contracts were very high. In order to generate income at a lower cost, the engineers had to put their heads together to come up with an idea. The plan was to focus their development efforts on a small set of open-source technology that could be sold. The key to maintaining this development on an order that could be handled by the group was to keep the focus very small. What they came up with was selling the GNU compiler (GCC) and debugger (GDB) as shrink-wrapped software. This was the right team of people to do the job. Michael Tiemann, who contributed numerous GNU compiler ports and also wrote the first native C++ compiler (GNU C++ or G++), took on the task of working on GCC; David Henkel-Wallace worked on the binary utilities (binutils) and the library; and John Gilmore worked on GDB.

This task grew to monumental proportions. One advantage, or so it seemed, was that John Gilmore decided to become the new GDB maintainer. Making this known to the Internet community immediately flooded him with different versions of GDB. Now came the task of integrating these new version features.

Eventually, the hard work paid off in what today is called the GNUPro Developers Kit. The kit includes:

Read the rest of the chapter yourself.

Re:QNX IS ON TEH SPOKE!!~1``

[AKAImBatman]

Wasn't that nice of moderators to moderate a polite comment out of existence? So, I'll repost it at +2 and if that gets modded to -1, I'll post it again. Here's the original response:

Here we have an example of the rarest of birds, an informed and insightful post... on slashdot.

QNX truly is the king of kings.

To which I replied:

Check my posting history. I don't always get it right, but I always attempt to present an informed and useful opinion.

And if you don't care for friendly banter, you can mark me as a foe. I'll get all broken up about it. Really. (rolls eyes)

Serious question here about end-of-life support

[grioghar]

Let's use Microsoft for an example, since they're so fun to pick on here on Slashdot.

Let's say MS was to release the source to Windows 98 4 years from now. Obviously no support for the software, MS is hoping everyone has upgraded. So, IMMEDIATELY people begin ripping the code apart, seeing blatant software security issues in the code.

Who is responsible for the ensuing chaos that results from the hacks and cracks that occur because now everyone knows where the buffer overflows are. Microsoft indemnifies themself, and then the user of the original software is left to hang.

I can think of one good example. I'm still a big fan of Quake II, and since the code release, there have been some SERIOUS cheats developed.

Just a thought