FreePBX: Victim of Own Success

[Lost Trunk]

Someone wrote in another thread that they'd like to see a new feature in PiaF. I had written the following before it dawned on me that there probably was a way for them to have what they wanted - a slightly clumsy method, but it would probably work - so I posted a relevant link and pulled back these comments (also, they would have hijacked that thread, which I didn't wish to do).

However, I think that those comments are relevant in the many cases where someone comes up with a good idea for a new feature, but it would best be implemented in FreePBX. But nowadays, it seems that getting a new feature in FreePBX is next to impossible (I won't even talk about the cases where I've heard that someone has handed them code on a silver platter, as it were, and it seemed as though they wouldn't even look at it - might be true, might not, and I don't know).

A couple of years ago, had someone asked for a new feature, I might have suggested that they submit it as a FreePBX feature request using their bug tracker.

However it seems as though most major projects reach a stage where they are no longer very interested in user input. Maybe the developers get burned out, or maybe they just want to move to another project. They start looking for reasons - any reason will do, no matter how ludicrous - to close a ticket (particularly if the first developer that happens upon the ticket is having a bad day). This is not always the case for true bug reports, but it definitely seems like they are no longer interested in adding features, no matter how reasonable the request may seem, nor how many people may want the feature.

I'm starting to fear that FreePBX may be a victim of its own success. Just as they are looking at an alternative to Asterisk (FreePBX 3.0 will support FreeSWITCH), ironically probably in part because the Asterisk developers have become increasingly difficult to deal with, I have a feeling that now would be a very good time for someone to develop an alternative to FreePBX, because they are starting to become as difficult to effectively communicate with as the Asterisk folks. And I actually think development of one or two competing projects might be taking place, although I can't recall offhand the names of any of the other projects.

Note that I'm not saying anything bad about PiaF - this is probably the most supportive community for FreePBX users out here (you're definitely more likely to get a helpful answer here than in the FreePBX forums nowadays). But I would hope that Ward would be flexible enough to move to another GUI based telephone software if it were better than FreePBX and Asterisk. Right now, as far as I know, there is nothing better, but if the FreePBX folks aren't careful their project is going to rot from within.

Just saying that maybe PiaF shouldn't be forever married to FreePBX and Asterisk. The PiaF concept is good, but both Asterisk and FreePBX are starting to get a bit long in the tooth (we're not talking decades here, but neither are brand new anymore) and the only way a project can stay successful over the long term is by being responsive to the needs of users, OR being the only viable alternative out there.

I know some will probably disagree (and if anyone from the FreePBX project reads this we'll probably get treated to a caustic reply) but anyone who's been around since the Asterisk@Home days probably realizes that the developers used to be a lot more approachable, and a lot more willing to consider feature requests (back in those days, they didn't even make you post them on the bug tracker - I remember one time someone made a suggestion on their IRC channel and "X-Rob" coded it right on the spot - sadly he's not a developer anymore). Things sure have changed, and not for the better!

 

[jroper]

Hi

I'm not sure that I fully agree with all your comments. I asked for a new feature, and it was coded up in a couple of days, however, I did pay to have it done, because it was important to my customer.

As to new features, look at the menu items in 2.5 compared to 2.6, there are many new features which have arrived, so I don't think that is a fair comment. When asking for a new feature in an OSS project, you either have to sell it by putting up a good case for the developer to put his free time into it, get it developed yourself, or pay FreePBX. All are valid options.

The FreePBX current framework lends itself to adding 3rd party modules, I cite the work being done by the team of programmers over a TShif's repository as an example, and just off the top of my head, can think of Concurrent calls report, endpoint manager and the very popular Superfecta as examples of new features and development.

In terms of other alternatives, I believe that this is a chicken and egg scenario. In AAH days, FreePBX was known as Amportal, and it barely worked at AAH 0.03 which was the first AAH install I did, but we persevered, and as a result, FreePBX is possibly the best there is, commercial or OSS. Whether it would have been the same success that we see today if Andrew Gillis had not rolled up an ISO so that we could easily do something useful within a short time of downloading is another matter.

As to other GUI projects out there, well I believe to bring them forward, they have to gain acceptance, and at least give the developers some encouragement, and that usually means rolling up a simple to install ISO, and advertising it's presence, but it will be hard up hill struggle, and possibly thankless, because inevitably it will be compared to FreePBX.

I've done an ISO in the past with asterisk GUI, and I've looked at Thirdlane as well, I've even downloaded and played with SIPX, but none were as intuitive as FreePBX.

FreePBX seems to enjoy a very good relationship with Asterisk and Digium, however the reasons for moving onto Freeswitch and other potential telephony engines are clear.

Freeswitch, developed by an ex Asterisk developer has got the benefit of hindsight, and as a result, is unlikely to have done the progammatical equivalent of painting himself into a corner, as one suspects is often the case with larger projects.

YATE as a telephony engine brings H323,, while the forks of SER, bring Nat traversal, and a whole host of other useful features, so to design a GUI that is independent of the underlying telephony engine is an obvious step forward to survive in the market place.

In terms of framework, when amportal was designed, there was really no viable OSS frameworks available, so you had to design your own. Now there are frameworks that allow you to quickly develop solutions, So to start again on a popular and competent framework would also appear to me as walking through an open door.

Finally, I'd prefer the FreePBX developers to be spending their time developing Version 3, rather than adding new features into Version 2.

Joe

 

[Lost Trunk]

Joe, just a few comments. Regarding being able to get a new feature but having to pay for it, I think maybe that's part of the problem. I don't think Andrew or Rob ever charged people to add a feature, though I may be wrong (if they did, it was probably because it was a rush job). The current developers seem to have a totally different mindset (something along the lines of, "how can we make money off of this thing?"). And in that respect, it seems to me that they don't honor those who came before them. Don't get me wrong, I don't begrudge them making money on their seminars and paid support, etc., but there a part of me that hates to see this turn from this cool project that people felt they could contribute to, to something run by people with more mercenary motivations.

Also, I didn't say that there are no new features in 2.6. But what I said (or at least meant to say) is that when someone submits a feature request via their bug tracker, my perception is that they rarely add it (even in cases where sample code has been submitted) - UNLESS it was submitted by another developer in their very small inner circle (and often that developer submits it first, then adds it, so I think they only do it as a reminder to themselves). Point is, if THEY decide they want a feature, it gets added. If a "plain old user" requests it, it's like they look for any reason to close the ticket and move on. If someone with a little coding experience suggests it and submits code, it never seems to be good enough for them, again unless it came from their "inner circle" - and it's not as though the process for creating a FreePBX module is exactly straightforward to begin with. TShif and a few others have managed to figure it out (and they certainly have my respect for that) but even then, their stuff gets relegated to the Third Party repository (which basically means that future versions of FreePBX can and probably will break those modules).

Basically, it's all about respect. The current FreePBX team seems to have an attitude that all the respect should be flowing in one direction, from users toward them, and that none has to flow in the other direction. I understand that as a practical matter, the person who develops software (especially free software) deserves more respect than those that simply use it, but when developers start thinking they can insulate yourself from users (and even others who might contribute code if they weren't made to feel their contributions weren't appreciated), and start to develop a "high and mighty" attitude, that's when the problems start.

Also as a practical matter, most anyone would rather see a feature they've requested rather than ten new features that are of no use whatsoever to them.

You mentioned SIPX, I think that was one of the ones I was trying to think of. I'm not surprised if you tell me they don't have as many features as FreePBX yet, but that doesn't mean they never will. And if they pay attention to FreePBX's weaknesses and make them their strengths, they could easily gain wider acceptance (same probably goes for other projects you mentioned). I think it would be healthy to have more than one project out there.

I will agree with one thing you said: "I'd prefer the FreePBX developers to be spending their time developing Version 3, rather than adding new features into Version 2." At this point in time, that would probably be the most logical way to go, and also, I get the sense that there is actually a more inclusive group of developers working on Version 3. Still, there are going to be a lot of people using FreePBX 2.x for quite some time (if only because it's rather painful to rip out an existing, working system and replace it). But I did notice that several folks seemed to have various issues after moving from 2.5 to 2.6 (things that had been working that broke after the upgrade) so maybe that's another indication that it's time to declare 2.x at "end of life" and move on.

 

[jroper]

Hi

it is not possible to argue with how you feel and your perceptions, simply to say that we would have to agree to disagree.

I would also make the point that a profitable OSS project, in my opinion, makes for a better quality product, and introduces an element of professionalism and commitment which is good for us as the users.

There is nothing mercenary about wishing to turn a profit, however there are lines in the sand that should not be crossed, but only your conscience, morals and beliefs, plus the howls of outrage from your users when you cross them, can tell you where those lines should be.

However, there are different groups of people to satisfy with OSS, and to make that point, I like the beekeeper analogy described here.

http://wiki.pentaho.com/download/attachments/8346/The+Beekeeper+(short)-V1.3.pdf?version=1

Joe

 

[wardmundy]

What gets icky in the beekeeper business is when the beekeepers start hoarding the good honey for paying customers and throw the scraps to everyone else. That's the concern that is being raised, I think.

I'm not suggesting this is the FreePBX, or Asterisk, or trixbox situation. But let's be up front about this. The primary goal of Bandwidth.com, Digium, and Fonality has more to do with generation of revenue rather than promotion of open source software. Only Digium made its commercial intentions for Asterisk known up front. You might also want to review the web sites of FreePBX over the past 3 years and compare the approaches. There's a good bit more in play than just a web site face lift. And it's worth remembering that all three of these companies derive substantial free honey from bees in their respective communities, and many of these bees aren't costing them a dime. :santab:

 

[jroper]

Ward

You are absolutely right, hence my comment:-

...however there are lines in the sand that should not be crossed, but only your conscience, morals and beliefs, plus the howls of outrage from your users when you cross them, can tell you where those lines should be.

Joe

 

[Lost Trunk]

Actually, the beekeeper analogy helps make the point of where things are falling down:

Bees can fly and so have the opportunity to leave the bee farm if they decide to. So the Beekeeper must tend to his bees. The Beekeeper has very little control over his bees and has no ability to direct them to do his bidding. Likewise the community can desert, or even worse fork, an open source project if they so desire (see the Wikipedia entry on Fork). So the COSS company must keep their community happy. The COSS company cannot rely on the community to follow any directive or schedule it might have. Bees can sting. Community members can publicly object or criticize the COSS company on its own or other web sites.
​

To put this another way, you have to keep the bees happy, or they might leave. They might get angry and leave, or they might just leave.

Each individual bee makes a small contribution to the system. It takes a large number of bees for a successful outcome. Likewise the contribution of the COSS company's community are vital to the model but the individual contribution of most community members is small.
​

Where it becomes a problem is where you have one or two, or even a handful of individuals trying to do everything. Not only do those people get burned out, but the rest of the bees feel their contributions aren't welcome (that is, they get no reward, not even a psychological lift, for their contributions).

(Backing up a bit in the document):

Each bee hive has a queen. In order to start a new hive the Beekeeper needs to attract a queen and enough of her bees to make the hive viable. Likewise open source projects often have a single founder or administrator that is the main leader of the project. Open source projects can be 'acquired' or 'merged' if the project leader and prime contributors are convinced that the move is beneficial to the project.
​

The problem is when the founder or administrator wants to be the queen AND the only worker bee (or one of a select few) that gets any recognition and respect. The analogy sort of falls down here, because in a real beehive the queen does virtually no work - she's the brains of the operation (seriously - if the queen dies, all the other bees stop working!). But in an OSS project, the lead developer is often one of the most proficient coders. We are not bees; we don't have one person with all the brains while everyone else slaves away at their assigned menial task. In an open source project, everyone who contributes has at least some intelligence, and some may even be as proficient at coding as the person who is nominally the head developer.

In a commercial open source project, those who contribute code (at least those not on the payroll) don't get rewarded in any other way than by recognition. Their reward is seeing that they've improved the project, either by their ideas and suggestions, or by actual contributed code. If the "inner core" of developers treat those contributions like big, stinking piles of dung, then the worker bees get NO reward at all, and eventually they fly away.

And that's when the developers sometimes get the attitude of "We're so under-appreciated, so we'll just do what we want (or what someone will pay us to do)." It's just karma - they're getting back the same vibes they've sent out to others - but sometimes they don't see that.

So that's all I have to say on the subject. I got interrupted in the middle of writing this, so I apologize if it's a bit disjointed.

 

[blanchae]

Now if only we could code in quantity and as fast as we can reply...

Discussion is good.

What I've found with an OSS project is that it starts out as someone who has a need, writes some code, others join in, the originator meets his need and surpasses it beyond his requirements. Then he is left administrating a project, long after he requires it.

I've tried various OSS content management systems and they just petered out due to the demands that the users place on the project leads. One day, you find that there is little support anymore.

In my own little world, I wrote a book 10 years ago, put it on line and it generates coffee money each week. There is little incentive for me to update as the effort is not worth the reward. Occasionally if I can justify an update based on what I'm currently doing, then I'll add it otherwise I look to what rewards me - that's human nature.

I have to agree that if you want features added within a reasonable amount of time then to use an old phrase "you have to put your money where your mouth is" otherwise those that reward will come first.

I'm sure that if Ward, Joe and myself weren't gaining financially in some manner, we would not be putting the effort in that we are. I'm fortunate in that as an educator, I get to be a mad scientist in the PiaF world without any real consequences as long as my students are learning and are happy.

I get to research and document how to use PiaF whenever I have free time which is a lot. I also have labs of Cisco switches and routers, DMS-100 and Meridian Norstar PBXs, ADSL, SONET, ATM and Frame Relay to play with. Life is good.

 

[cosmicwombat]

I have a different view.

As one labeled a "Contributor" (read as not a Developer) but none the less involved with FreePBX, I have frequent contact with the developers and see ideas and concepts bantered about by them all and fairly consistently.

As phone systems are different in business use then say a desktop computer, you plan migrations and upgrades as infrequently as possible. It could be a good while before V3 is as usable with FreeSwitch as V2 is with Asterisk today.

While those willing and able to make most of their configuration changes in CLI are running V3 in production today, I think it is safe to day that most of the current FreePBX crowd would rather configure most things via the GUI. I know I do. So we are talking a year or more.

My approach to all this is patience. I was an AAH newbie and have been involved ever since and don't think things are nearly as nefarious as sometimes made out.

Also, please don't use phrases like [sic]"(and if anyone from the FreePBX project reads this we'll probably get treated to a caustic reply)"

I just don't think hostility is productive.

 

[Alex728]

another issue is that the perceived value of the features may differ to each group of users. There are a lot of people here who are very talented engineers and understandably want the "latest and greatest" and to push the boundaries of what can be done with a telephone system, or people who are resellers and want a USP to distinguish their system from the competition.

but there are also those who are perfectly happy with the current featureset of something like FreePBX version 2.x - and are also deploying a VOIP PABX to co-exist with analogue circuits and deliver more at a slightly cheaper cost than something like a NEC XN120 or Panasonic system (which PiAF and other distributions do admirably).

I can't (yet) code entire modules but I am also trying to use spare time to learn more, and what I do learn (such as the quirks of the British analogue telephone network over the years and how to get it to "play nice" with Asterisk) I am always happy to share.

I first experimented with Trixbox in 2007 and the features there were more powerful than anything my organisation used previously. I switched to PIAF not so much because of politics but more as I found I needed to hack at some of the source code to get it deliver consistent and reliable caller ID from analogue telephone circuits.

Now I have 3 PIAF boxes in a linked numbering scheme across 3 sites, all running FreePBX 2.5 with the latest modules and they are rock solid. When things "go wrong" with them its invariably due to British Telecom line faults or the electricity supply cutting out in the night long enough for the UPS batteries to be exhausted. These things happen in a rural area, and I have appropriate contingency plans.

As an aside I have an interest in nature and the environment - and there is a real-life shortage of bees!

 

[Sippy]

Asterisk/Freepbx/Linux/Windoze/MCE/SIPX/SIPSwitch/

Little ole me has watched all of these projects for the past few years and each one has benefits that all of us could learn. My mantra for all of these systems has been when is someone going to make a phone system that is reliable.

Any customer or user will demand this after a few weeks of problems that keep the system down or broken. PIAF in my opinion is the only salvation in the wild wild west of VOIP.

If pure SIP is the answer then a PBX is not needed. There is a ton of companies offering a SIP phone and service for $40.00 per month. If you want to have a centralized system that can record every call, or do a conference bridge then PIAF is the answer.

You can always throw money at a problem and solve it. It takes heart and soul to do it with a passion to make it work. That is what open source should be. I am very thankful for the folks that help this project. They do have the heart and soul.

Just my two cents.

 

[mickecarlsson]

I am replying to the first poster to give an update on how FreePBX is moving forward but first something about me.

I am one of the newly added FreePBX developers. I have worked with FreePBX for more than three year now. First as a user, then as a contributer and finally as a developer.
We use FreePBX/Asterisk at my work and have been doing that for more than three year. We have an Asterisk BE license from Digum but went with the asterisk.org version together with FreePBX.
When I initially started with FreePBX I found some bugs and a lot of spelling errors, I offered them to the FreePBX community as diff files but was granted access to svn so that I could submit them as soon as I found them.

I also submitted a couple of feature requests that was granted. I then started to work through the bugs listed in the tracker to see if I could do some work on them.

I worked closely with the rest of the development team for the 2.6 release and here are some facts about it:

There was 173 Feature Requests made and 130 of those were implemented.
The rest was flagged as worksforme, duplicate, invalid or wontfix.
There was 317 bugs submitted and 180 of those were fixed.
The rest was flagged as worksforme, duplicate, invalid or wontfix.

That is why we use the bug tracker, just to keep track of things.

To compare with 2.5:
There was 178 Feature Requests made and 122 of those were implemented.
The rest was flagged as worksforme, duplicate, invalid or wontfix.
There was 385 bugs submitted, 248 were fixed.
The rest was flagged as worksforme, duplicate, invalid or wontfix.

I feel that we listen and try our very best to make FreePBX better and better.

They start looking for reasons - any reason will do, no matter how ludicrous - to close a ticket (particularly if the first developer that happens upon the ticket is having a bad day).

Well, I suspect that you was looking at one or two of those tickets that was created by a user that did not agree with me closing his ticket(s). These incidents made me very uncomfortable and I was seriously thinking about leaving FreePBX for good as I received many nasty PM's and accusations. The solution to all this was that the user took a timeout. End of story as far as I am concerned.

I'm starting to fear that FreePBX may be a victim of its own success. Just as they are looking at an alternative to Asterisk (FreePBX 3.0 will support FreeSWITCH), ironically probably in part because the Asterisk developers have become increasingly difficult to deal with, I have a feeling that now would be a very good time for someone to develop an alternative to FreePBX, because they are starting to become as difficult to effectively communicate with as the Asterisk folks.

Well, FreePBX V3 is not just for FreeSWITCH, it is also for Asterisk and others to come. At first it was for FreeSWITCH as the GUI was developed for that, but the new developers that work on V3 is making V3 independent of whatever engine you use.

I don't recall if I have spoken or mailed you Lost Trunk, but if you feel like it, give me a PM or meet me at #freepbx-dev on IRC were I am hanging out.

I have made one presentation about me on FreePBX, read it here freepbx.org/news/2008-10-28/an-introduction-and-an-announcement

Mikael Carlsson
FreePBX Development Team

 

[wardmundy]

Thanks for the response, Mikael. For those of you that don't know Mikael, please read his blog posting. His work on the FreePBX language module was and continues to be Herculean!!

 

[blanchae]

Great to hear from you Mikael, I've read your patient responses to some of the FreePBX forum posts, it's good to see that you still have time to help the new guys.