Topic: Future of SME

[PhilV]

Hi, I am just wondering what of the future for SME. I was about to implement an SME server at a company, but I am now not sure whether to proceed or not?

I understand that SME is bassed on the Redhat kernal, and that is now no longer open-source as such, so where does SME go from here. Will it be stuck with nowhere to go, and gradually become outdated, and therefore open to new security hacks etc.?

I am wondering if anyone else has wondered this, and if they have reviewed any other software package like SME. This is the first I have looked at, but would be interested to know what alternatives are out there, (for free!), that I could also look into, just to be sure that i have the best distro for this application.

Many thanks,

Phil

[cydonia]

Me too... I'm beginning to use SME more and more, but getting a bit wary now.

ClarkConnect offers similar features.  I love sme, but i wont stay with something that just gradually outdates itself and becomes obsolete.

[PhilV]

Anyone else got any comments?

[Jon_Reynolds]

I would think that with the SME server being open sourced that it will migrate to whatever OS will support it. Just because it is based on RH doesn't mean that it can't run easily on another Linux type OS. Maybe the best route to go would be to use LFS(LinuxFromScratch). I don't know, just speculating.

Jon

[Jon_Reynolds]

I would think that with the SME server being open sourced that it will migrate to whatever OS will support it. Just because it is based on RH doesn't mean that it can't run easily on another Linux type OS. Maybe the best route to go would be to use LFS(LinuxFromScratch). I don't know, just speculating.

Jon

[Anonymous]

Phil,
   I share your concerns about the longevity of SME based on old RH7 code.  I am equally concerned with the management of SME by contribs.org.  I realize that this is a big undertaking and I appreciate Jeff, Hsing and others efforts, but things appear not to be turning out well.  I am concerned because of the lack of mailing list activity, the recent outage, and unanswered concerns from the Mitel->contribs.org transition.  
   IMO, the following needs to happen to ensure SME longevity.

1) Move the bulk of SME to sorceforge or equivalent to save expenses.  This includes files, documentation, mailing lists, bug tracking, cvs, etc.

2) Move SME base from RH to Debian so security updates are less expensive and more reliable.

3) Setup a nonprofit org to fund #2 above and to define the SME distro goals and future direction in an open fashion.

Thanks,

Dale

[dalesyk]

Phil,
   The previous guest comment belongs to me.  I was logged out by accident when I posted it.

Dale

[bobk]

The following is my personal opinion and is in no way meant to belittle the efforts of the contribs.org staff and volunteers. Every one of them seems to be working very hard. However, I do feel that their efforts are being applied in the wrong direction.

Contribs.org took over the core distribution from Mitel in December 2003. Since that time there seems to have been no visible development work on the core distribution. Excepting the 'face-lift' release in March 2004. There is still no published development plan or even a general road map of future direction or development. It appears that all the available resources are being applied to produce and maintain a fancy web site. This web site appears to be directed at just what the name says "Contribs". There is very little real information available about the core distribution and the little info that is available is difficult to find and access. If I were a potential new user, I would be put off by the lack of easy to fine basic information about the distribution itself. Documentation of contributions seems to be progressing but there is an almost total lack of visible effort being placed on core distribution documentation.

I personally believe that the core distribution needs to be split off from contribs.org and set-up as a true open source development project. Contribs.org should continue in the direction its name implies and support add-on contributions to the core distribution. As an open source project, the core distribution should not be connected in any way to any other organization. It needs a dedicated group of volunteer lead and contributing developers that are not concerned with web site development and/or add-on contributions but concentrate solely on security and improvements to the core distribution. It needs a clearly defined and public Project Development Plan and accurate up to date user documentation.

I could go on and on but I think this expresses my basic opinion. Again, I am not trying to belittle the efforts that have and are currently being applied. However, without a solid, secure, up to date and progressing core distribution what need is there for add-on contributions.

[duncan]

bobk - I think that is one of the saner suggestions floating around here. Split the core off and contribs can do as they like with the rest.

[tape]

The future of SME is.. you!

Contribs.org or Sourceforge? It does no matter where SME will be hosted, and the guys from contribs.org never said anything different than they just try to help as much they can.

They also always told to be open for any help or suggestions, but what if there are none?

And, SME of course is Open Source, everyone can take the code and build his/her own distribution.

But, is complaining better than helping? Does complaining change things better/faster than just doing them?

I am no programmer (i am running a SME server in my intranet), and i am helping with some small amount of money. Just as i personally can do, nothing more or less.

Also i am here in this forum as a registered member, trying to answer questions or ask some if needed. This is how i am helping, what about you?

The future of SME is me.. and you. Nothing else! And evereyone in his own way!


best regards from South Germany
Anton Pirnat

[Anonymous]

I don't think that anyone here is trying to put the blame for the lacking development on the people behind contribs.org, if they are then shame on them.

When Mittel gave us all E-smith they also gave us a task we was not up fore. Before the project can move on we all have to agree on the goal. That alone can be a difficult task since there is just about as many opinions about the goals as there is members in this forum. When Mittel owned E-smith it was much easier. They decided what to do and when to do it.

So far all we have done is carring on the forums from e-smith.org. What we need to do is have an election about the future. After that we can make groups, which can take care of the development their subsection of SME.

In the mean time let us make sure we still have a forum to grow our beloved SME.

/Carsten

[lajgaard]

It's my post above. I was logout before I hit "submit".

[schrambo]

Firstly a plan for future development of SME needs to be created, as in what direction SME will take off. At the moment SME is standing lost in a desert without a compass or map. We need to take acton and decide on what the majority of the users want the future of SME to become and do it.

Personally I would love to see SME ditch the RedHat chassi and adopt either a slackware or debian framework just to name one of many changes I would personally make too SME. Sadly its only in my wildest dreams that I could do such things. I would very much love to help in the future developments of SME eithe it be testing, user suggestions and documentation.

SME could and should be adopted by its community. I think having the development of SME seperate to contribs.org as metoned by others would be a more ideal system, not having to rely on the contribs people to continue development of SME but maintence of the contribs aswell..

[gelcube]

It seems to me that the thing that brought most people to SME was it's ease-of-use and easy administration.  So a majority of SME users are just that...users like me.  Not programmers.  There should be a list of programmers who are willing and able to work on SME.  This list will become the "Core Team".  That team will be responsible for the official SME development.  Right now, there's no core team to work on and release new SME distributions.  

Of course, that means that some of our community are going to have to step up to the plate.  Who here can program well enough and are willing to work on the SME kernel?  And who wants to be responsible for the development roadmap?  Anyone?

[MSmith]

I've been using the e-smith distribution since version 4.1.2 for various things, and it's great.  Was great, is great, has a great future.  Except, of course, for the cold fact that development takes time, energy and programming skill.  Which are obviously present ... but in sufficient quantity?  Who knows?

When Mitel set the distro loose late last year, that represented an enormous challenge to anyone who might want to go forward with it.  I mean, put yourself in the new developers' shoes.  You have to familiarize yourself with a bazillion lines of source code, put a team together and *modify* this fancy piece of equipment ... taking into account the fact that it'll be routinely exposed to the harshest Internet environment ever known.  Hack attempts, worms, DOS ... so if you're the developers you feel this responsibility keenly.

Not only do you have to deal with outside threats ... you have to deal with Microsoft's changing the rules routinely so your Samba has to be up to date ... and you have to maintain backward compatibility with Mac OS 9 & 8 while supporting OS X ... and PHP and MySQL keep on advancing.  

Plus there's this user community ... all of whom (myself included) would love to see this or that feature but disavow any programming knowledge or time to do anything meaningful.  Add to that the constant parade of those who Just Don't Get It who want X added to a distribution from which it was deliberately removed ... it's enough to drive a person to drink.

So maybe it's time for the sunset of SME.  Either that, or some real traction in the direction of a nonprofit foundation to keep the code in trust while developers come & go.  I'm up for that, and would support it with $$$.

P.S. It WOULD be nice to see that moribund Sourceforge listing actually put to use ...

http://sourceforge.net/projects/e-smith/

Looks kinda sad that way.  It'd sure attract worldwide attention if it were up to date.

Then there's that ancient and decrepit Freshmeat listing:

http://freshmeat.net/projects/e-smithserverandgateway/

[Anonymous]

Some of you guys have a point.

However, you are all also missing a point.

Every company, no matter what, where, when,
or how must have a single leader in charge.

When it comes to making the decision as to what direction that SME should go in, there can be no we. It has to be an I as in I am going in this direction, if you agree follow me. This of course will offend many, howeveer, it will give concrete decisive direction to the project. So as opposed to voting on a direction for the future of SME, why not vote for a project director who has a vision and a direction that he or she would like to see the system go in?

A community can provide suggestions, however, there can be only one person with the final decision and responsibilty to lead and direct the project. This person would also be responsible for providing a monthly report to the community as to the status of the project. The project director would be primarily responsible for testing the changes and thereby making a decision as to whether a development is stable for release or not.

This is how Linus Travoldi develops the kernel. It works for the kernel and it can work for the SME development project as well.

Once a project developer has been chosen, he can set up a non-profit organization to receive donations and membership dues. Where Only due paying members would be allowed to vote for the board of directors and the project director. Requirements for President and board members would be as follows. He or she would have to be first and foremost, a due paying member, other eleigibilty requirements could be set such as, achieving a certain amount of time invested in the project. The criteria for determining this would be based upon area's such as developement, contribs donated to the community, support given through the forums ect. In this way, only those people who actively give to the community and support the project in one way or another will be eligible to run for the position of board member or the president (project director) of the corporation. The dues paid by voting members would then be utilized to pay developement costs such as salaries for the President and board members, developers, and those additional staff members that would be needed to support the project.

While tis wouldn't allow for every little things someone wants added to the system, this would allow for a proper and continueing development of the SME distro, as well as maintain it's free distro standing.

Scott

[Anonymous]

Why does this things post as guess when you are logged in? I posted the previouse message.

Satovey

[satovey]

relogged in just to make sure it uses my username this time.

[satovey]

relogged in just to make sure it uses my username this time.

That is just wierd, I wonder why or how I logged out?
nuff said.

Scott

[Anonymous]

Red Hat Linux (RHL) has a new name, Fedora Core (FC).  This was about opening up development as much as the trademark issues.  And an entire new, Debian-like distribution model was built around it.

The private RHL lists became public Fedora ones.
The RHL developers started working on Fedora Core.
Red Hat opened up its build systems to non-Core packages.
Red Hat opened up its control to an Apache Foundation-like steering committee.

Fedora Legacy (FL) continues to crank out RHL 7.x, 8 and 9 updates.  It will do so when FC 1 is EOL very shortly.  The standard policy is about 18 months after EOL -- about 2-2.5 years after release.  Hmmm, that sounds a lot like what we had with RHL (which was _never_ guaranteed either!).  Not surprisingly, there are a _lot_ of paid Red Hat people working on Fedora during Red Hat time -- and not because they are doing it "behind managements back."

Red Hat has _always_ valued a 100% GPL-centric focus.  Be it the 100% GPL'ing of multi-million dollar packages gone commercial (I hope _everyone_ caught the recent GFS stuff , to 100% GPL'ing their _entire_ Enterprise line.  You want RHEL?  You can get it, for free, although unsupported.

Fedora has shocked me.  Red Hat did good.  People say Red Hat is evil because they introduced Red Hat Enterprise Linux (RHEL) and dropped RHL.  They did _not_ drop RHL, they made it better, and now call it Fedora.

SuSE does _not_ certify non-Enterprise flavors either.  And what most poeple don't know is that SuSE was the _first_ distributor to introduce the Enterprise product.  When Red Hat's enterprise customers said they would leave for SuSE unless Red Hat did they same, they introduced it.

But instead of leaving a split shrink-wrapped model, like SuSE, Red Hat decided to open RHL like many of the RHL developers themselves have _always_wanted_!  With the introduction of RHEL, there was no need to keep the chains on RHL, and RHL was opened up.  The name change was because of the _real_ trademark issues with doing so.

I love FC and will continue to trust it for production systems, just like RHL before it.  At the same time, thanx to Fedora Extras, Legacy and 3rd Party, I have a wealth of software and updates just an "apt-get install" away.  Best of both worlds.  Again, Red Hat did good.  And they continue to put _paid_ people on FC, just like RHL before it -- just like they do on GPL projects MORE THAN ANY OTHER DISTRIBUTOR!

[ P.S.  Don't forget what RHEL $$$ go to fund, only more GPL software!  Find another, non-GPL vendor to spread rhetoric about. ]

Bryan J. Smith
b.j.smith@ieee.org

[Anonymous]

I have _nothing_ against Debian.  I used to produce .deb packages myself for a major Linux project (as well as RPMs).

But since e-Smith/SME Server is based on Red Hat Linux (RHL), Fedora Core (FC) and other Fedora packages (like Extras and Legacy) would be my recommendation.  Just like Debian, you can apt-get official packages from the Fedora projects and repositories, as well as 3rd party ones too.

Or we could consider something like Progeny's Componentized Linux which aims to leverage _both_ Debian _and_ Fedora.

I'd advise against Linux From Scratch (LFS), too much headache.  Gentoo and Slackware are not bad distros, but I really have a long-standing preference for the stability of Red Hat kernels.  That's just my opinion.

I love the stability of Debian too, and their releases are just as well integration tested as Red Hat.  Unfortunately, I find that Debian development and release takes to long, and I've really hit some whoppers mixing in testing in the past.

Red Hat's direct involvement with Fedora Core (FC) keeps it being released with the best balance of time-to-market v. quality assurance IMHO, just as much as RHL before it.  Until I see the quality in FC slip, I don't really see any reason to change it if porting to a new Fedora version is found to be easiest.

Remember, FC=RHL.  Even Red Hat's attitude on changing FC reflects this -- i.e., they don't rip things out for a reason, hence the 4 CDs now -- yikes!  But at least Fedora-based projects like Cobind show it can be reduced with proper focus.  I would like to do the same with a 200-300MB "common denominator" base I call "Quark."

Using Apt (or Yum), there is _little_ reason to ship a new, Fedora Core-based SME server that doesn't fit on a single CD.  You can always apt-get (or yum) anything else you need, as well as updates, without having to directly support the infrastructure that does it.

Bryan J. Smith
b.j.smith@ieee.org

P.S.  In case I didn't make myself clear, I'm all for Debian too!  I was just saying if it's easier to move to FC from RHL (which it should be by far), I'd recommend that course first.

[Anonymous]

Hi Bryan. Are you aware of the start being made here.

http://lordsfam.net/sme70

[Anonymous]

This looks good... maybe the community should get behind this guy? Someone who has stepped into the breach, side stepped all the waffling, and actualyl got on and done something?

Phil

[satovey]

This looks good... maybe the community should get behind this guy? Someone who has stepped into the breach, side stepped all the waffling, and actualyl got on and done something?

Phil

  I posted a similer statement in another forum, however, it seems the community wants to keep discussing which direction SME should go. I think that since slords has taken the initiative to port SME to the Fedora Core, all debate should cease at this point. While debian sounds to be a good alternative, no one with the talent and skill to port SME has begun to do so. Therefore, I say that Fedora has won out by default.

Scott

[hdaei]

Great to see this updated distro!

Like many here I'm only a user / admin, not a programmer...  Thank you http://lordsfam.net/sme70 for going for it.  If I can help please let me know how... I can do a little beta testing if need be.

Douglas

[tape]

Great to see this updated distro!

yup.. same here, it really looks interesting! Using FC2 would be a great way (imho)!!

It is experimental with many bugs or working with small needed changes? Should we wait or test it?

Anton

[duncan]

At the moment it looks like Shad is upgrading an existing 6.0.1 release with Fedora packages.

I doesnt look as if there is a "Distro" as of yet.

[azche24]

Smoke free since March 2004

Stick with that, Duncan! It saved my life to stop smoking > 30 cigs a day and i feel better every day after more than 2 years now.

Also you are right in the topic: The core should be split from the contribs. The contribs could stay somewhere, e.g. at contribs.org - but:

Who will develop the core? I am user and no programmer, i have no time at all and will never be able to do any developement in this  

So someone has to do a little financing and/or fundraising. Or the folks, that make their living out of installing an customizing e-smith (the many consultants around here) should particapate a little more in the core develompement. They already do in the forums and in the support side of the thing.

Contribs remain add-ons but someone has to take care of the core-developement.

[duncan]

Off topic I know but.

Thank you Alexander. At four months the pain is gone - and I am amazed at how much better I feel. Never again.

If you take a look over on the mailing lists http://no.longer.valid/phpwiki/index.php/DevList it looks as though there is some movement towards maintaining the core.

Regards Duncan

[rmccue]

One thing we might consider doing is using http://WhiteBoxLinux.org as the base distribution for the SME server.  From their web site:

"This product is derived from the Free/Open Source Software made available by Red Hat, Inc but IS NOT produced, maintained or supported by Red Hat. Specifically, this product is forked from the source code for Red Hat's _Red Hat Enterprise Linux 3_ product under the terms and conditions of it's EULA."

The advantages of using WhiteBox is that it is compatalbe with Redhat Enterprise Linux 3, and will have patches availble for a much longer period of time after new releases come out than Fedora.

Just an Idea.

[Anonymous]

A thought -- not a demand -- not a rant -- not even really a suggestion...

But, if the development platform migrated to Debian, it might make for a more diverse user base.

The reason I say this is that esmith was the first distribution I used to set up a Linux firewall/gateway router in, I believe, 1999, after paying for media and documentation, but no support.

It was great, and one of the great things about it was that you could, with some workarounds, size the hardware for whatever task you wanted to deploy emsith for.  After setting up a P120 at work, I used esmith on an ISA 486 and an external modem at home (there were workarounds for ISA motherboards, etc.).

By switching to Debian, this hardware compatibility would all be opened up again. Ease of install obviously requires the new user to stick with PCI motherboards, and reliability issues might dictate leaving the venerable 486 behind, but by exploiting Debian's lower machine requirements, the distribution would open the doors to more trial, experimentation, and use.

A Debian based server could potentially make a Pentium 75 with 16 MB RAM a viable option (the old box I used to run esmith on was a 486DX 66 MHz with 16 MB and a 500 MB hard drive). For many people, curious to try a distribution, it is easier to come up with something like that, than to scrounge up a dedicated test box that meets, for example, Fedora Core2 hardware requirements.

This would also make the distribution more competitive with "firewall only" projects, for those who are interested primarily in firewalling and NAT, while not diminishing its value to those looking for a full blown solution for a small business with 150 users.

Again, just a thought and no demand, but migrating to Debian would make it all more accessible to a broad cross section ranging from the home user to the 5 seat third world startup to the well established 500 employee small manufacturing firm in New Jersey.  In addition, it would lower the bar to participation by a larger cross section of developers and bug testers.

Colin Mattoon
Lewiston, Idaho (U.S.A.)

[azche24]

In addition, it would lower the bar to participation by a larger cross section of developers and bug testers.

Yes, Colin,

though i am not experienced with debian: If debian would be able to handle even low-end hardware (if!) it always would be my choice instead of RH.

I liked linux first for little hardware demands. It just was faster than win on my low-end PCs, which we also use for achieving low energy consumption and less heat in an office environment. E-Smith is still much more safe than a similar windoze server. But the smashing argument for linux servers (besides the license fees) will always be a low-end hardware approach.

And even my home-server now needs 256 MB RAM + more CPU-Power than this 350 MHZ PII PC from 1998 can give....

[EnglishRob]

I like the idea of using RPMs myself.  I'm not an expert on Linux, I've played around with it on and off, but I like the idea with SME Server that it just works.

My SME server is running version 6 IIRC, and it is working nicely on an Athlon 800 with 256MB ram.

With hardware being so cheap nowadays, would having to support anything lower than say a P2-300 with 64MB or so memory really be a big issue?

I've bought components like P2 CPU's, motherboards and memory cheaply enough on eBay and even low end Duron's aren't that expensive anymore, or for the low power option whats wrong with the VIA Mini ITX motherboards?

Just my 2p.

Rob

[EnglishRob]

Just one niggly problem I find with SME server at the moment is creating template files.

I like the web interface, but I would love a nice easy way of creating/editing these template files.

That is, unless there is any easy to understand howto's on doing out there?

Rob

[wellsi]

Just one niggly problem I find with SME server at the moment is creating template files.

I like the web interface, but I would love a nice easy way of creating/editing these template files.

That is, unless there is any easy to understand howto's on doing out there?

Rob

Although straying off the subject for this thread...

What problems do you have with the template files? Is it the concept or the actual syntax?

Some constructive criticism is this area would be good so that we can improve the documentation.

I think that there are no HowTos however there are the pages in the documentation section.

[EnglishRob]

Okay, I think its the actual concept of it all, or maybe I'm just that I can't get my head around Perl?

I don't have many problems with RedHat/Fedora etc with their config files, it seems to be the same as pretty much every other Linux I have come across.

I kinda see the advantage of having the config files separated into chunks so just adding a file rather than editing config files gives the same results.

I just wish there was some sort of editor of sorts which allowed me to create files with all the nessicary bits in it without having to read through other config files to try and understand how they work.

I must admit, I probably should spend a bit more time looking at them, I just can't seem to find the documentation on them, and what I did have (I think was in a PDF), wasn't too clear to me.

All in all though, SME server works nicely for what I use it for, and its even better considering there aren't any hefty licence costs.

Rob