Upgrading to v4... or may be not.

[visionit]

Hello Folks

It's been nearly 2 years since I've posted here and the time has come to update the helpdesk, along with other related software.

We've had a quick look around the new v4 demo and apart from the obvious problems (PHP warning messages fill every page, something to do with array keys & the Zend engine)... I have to say I'm very surprised at the changes.

I know the team have put an awful lot of hard work in to version 4, but I can't help but prefer the previous version... it felt more feature-rich & rounded somehow.

The customer ticketing interface (to be brutally honest) looks a little like a front-page effort by a newbie. The old interface was (and still is) one of the best I've ever used. The admin interface has lost the fundamental functionality of a true helpdesk (SLA, time-management etc)

OK, I've heard there's a "plugins" section but you shouldn't have to "plug-in" any key parts of a helpdesk. Plugins are great for end-user additions but it shouldn't be used to bolt-in scripts which should reside in the core of CH.

I hate to say this (as I've always loved the Cerberus platform) but it's no longer a viable choice for us and we're forced to migrate to Kayako. Yes, it's £300 and only supported for 6 months, but it has all the functionality of the old v3 cerberus platform with the added bonus of native PDA support.

I don't want you to think this is a kvetch thread because I have nothing but praise for the continued efforts of the team... but if I (a genuine cerberus enthusiast) has look elsewhere, what hope is there for newbies?

If Cerberus is ever restored to its former glory, I'll have no compunction to drop Kayako and move back.

A re-write is good, but only if the end result is more efficient and as functional as its predecessor. What's there is great, but the immediate impression I get is v4 is an un-finished (though is anything ever finished?) engineers play thing, not a well-rounded open-source rival for other similar helpdesk packages.

Very sorry guys.

[xps]

Its said but i must second most of what you said in here.
I was a very happy customer and also did something on the beta list and stuff like that.

Sadly i think Cerberus has taken the wrong turn somewhere. Lots and i really mean lots of the function i would expect from a ticket system has been gone. I cant see a reason why this has happend.

Wheres the support-center for example? (and please dont try to sell me the current one as a serious alternative to the old one which wins both hands down).
Wheres the Workstation client?
Wheres the Createt Ticket link?

And thats only some things im missing.

And cerberus guys, have you taken a look at your demo lately? How you think you can sell this script with a demo looking like that?

Its really sad that cerberus is no longer an option for us. We stick with 3.x until our new system is running and than move away with some good experience from 2.x and 3.x.
If anyone is interested were moving on to saeven|crm (former auracle) which has every little of those options build in and not as a either not existing plug-in nor a plugin that looks like a childish version of what it was in 3.x.
It also has an actually working live-chat and a workstation.

Im very sorry as well although i really wish you all the best for the future!

[visionit]

Its said but i must second most of what you said in here.
I was a very happy customer and also did something on the beta list and stuff like that.


I still am a happy customer. v2.7.0 has served us VERY well and my disappointment in v4 in no way affects how truly grateful I am that v2 allowed the business to grow with a reliable and outstanding support desk. We developed many plugins (they were known as "modifications" in the old forum) which greatly improved and streamlined daily tasks (anyone remember Cerberus QTM?)

Sadly i think Cerberus has taken the wrong turn somewhere. Lots and i really mean lots of the function i would expect from a ticket system has been gone. I cant see a reason why this has happend.

I have to agree. I think the decision to start-a-fresh was a bad one. I can see an awful lot of functionality has gone, but there are almost daily updates by the team... and v4 hasn't (from what I gather) ever been advertised as a turn-key solution, or finished for that matter.

Wheres the support-center for example? (and please dont try to sell me the current one as a serious alternative to the old one which wins both hands down).
Wheres the Workstation client?
Wheres the Createt Ticket link?

I miss the support-center the most to be honest. This is our current setup (hasn't changed in over 2 years) http://helpdesk.e-visionit.co.uk . Hit "home" and you'll see the helpdesk and site were skinned to match perfectly. The inability to "create a new ticket" isn't just a post-beta missing feature... it's absolutely bizarre!

And cerberus guys, have you taken a look at your demo lately? How you think you can sell this script with a demo looking like that?

I suspect that's a recent error, but I agree it doesn't look good. I ignored it because I understand why it's there, but other users will be put off by it.

Its really sad that cerberus is no longer an option for us. We stick with 3.x until our new system is running and than move away with some good experience from 2.x and 3.x.Im very sorry as well although i really wish you all the best for the future!

I would love to stick to v2.7.0 but recent events have forced us to review our current systems (not just the helpdesk) and lead to Cerberus v4 (I believed the natural upgrade).

The old system may have been cluttered and a little belts-&-braces in places, but it worked. OK, you have to take a different approach with open-source software but businesses just aren't interested in how & why these functions aren't available. It's really quite simple...

Does it do the job we want?

Yes/No

If the answer is no (regardless of the reason), people WILL look elsewhere.

Please guys, I'd urge you to rethink your strategy. Tune & fettle the previous version, add a "plugins" section to that... make it quicker/more efficient in key areas (parser, XML configs etc) but don't continue down this route.

I love(d) Cerberus because of what it was & what it was capable of (and the support of a dedicated team of pro's) but you've taken all that away. Take off the "cerberus" badge and if I didn't know any better, I'd say someone has tried to copy a formely superb platform and butchered it in the process.

[jstanden]

I'll respond to both of your specific points in detail shortly, but first I wanted to highlight something fundamental that you both are overlooking.

The core of our renewed development ethos is to share our work with the community as soon as possible.

Logically:
If we include every feature request by default, then the project turns into an incoherent mess which isn't usable for anyone.
If we only develop the features we personally need, then this ceases to be a shared project worthy of marketing to other people. Ergo, there is no reason for us to be in "stealth mode" during active development.The clear path for us is to use our experience in writing software to be the full-time custodians of the project, while working closely with the community to determine what we're actually building. We are making as few assumptions as possible before handing over something tangible to you guys to elaborate on with us. Without actually seeing something working, many people don't know what they actually want. That's just how software development works. But it means on the path to the best answer, we're sometimes putting out tentative "feelers" to provoke feedback.

This long-term approach is powerful, but it's significantly different than how things were done in the past, where we'd announce a major release after a long period of silence and the result was just "how things are now". I can't imagine you'd want to return to that.

A major problem, highlighted by this thread, is that you are carrying on that outdated mindset of assuming what you see is the completely finished solution, meant to solve every problem or descend seamlessly from every prior feature. It's not. And we've been very clear in our communication, from the moment we introduced 4.0 in beta, that our goal is to rebuild Cerb4 from the ground up with the community able to provide feedback every step of the way. This isn't a box of immutable software sitting on a shelf somewhere; it's an agile, interactive project which makes daily progress.

What we have right now works incredibly well for the realistic, fundamental needs of everyone. We have a freshly-engineered platform that makes incorporating feedback very painless, and we have a re-energized development team that's telling you anything is possible. It's silly to underestimate the impact both of those factors will have on development going forward.

The small luxuries on top of this platform, like advanced SLAs and time-tracking, are absolutely trivial to implement from a development standpoint.

If we knew exactly what we wanted to build for such features, they would be done in a day or two from the technical-end. However, realize, at WebGroup Media we don't use time-tracking or advanced SLAs. We're very experienced software developers, but we're honest enough to say we aren't call center consultants. There are genuine experts in our community who know exactly what a feature like time-tracking should encompass. It's much smarter to ask for their feedback than for our developers to fake it (again). Agreed?

And it's a much better 'check & balance' to expect someone requesting something like time-tracking could provide a fairly detailed explanation of how they intend to use the feature. If they can't, we can generally assume they're just shopping from a dispassionate list of marketing-driven bullet points. There's no sense in hacking away on the project to appease such things.

On a purely objective level, I agree with the gist of what you guys are saying. As a fellow business-owner, I generally don't have the patience to let open-source democracy take its course from a completely clean slate. I look for complete solutions that I can plug in to solve a problem with minimal time or cash wasted.

But over the ensuing months and years, as I constantly work around the shortcomings in these prepackaged solutions, I realize that the main feature I'm missing is a responsive, transparent, personable development team who embraces change because it's inevitable. That is exactly what we offer you.

On a very short timescale (weeks), focused myopically on completed features and nothing else (which, I remind you, are technical trivialities), you may very well feel better served today by one of the alternatives you mentioned -- be it a competitor or Cerb3. But, realistically, some of the alternatives you have mentioned fail horribly in the long-term measures of developer transparency; including "full-attention" dialogue like this, source code availability, support, etc.

Cerb4 development is exponential. By treating every development task as an opportunity to expand the toolset, new features become cheaper and faster to develop than early features -- instead of the exact opposite, where development slows down the longer a project is active.

You know what you want for your own universe. At the time same, for a project serving as many groups as Cerberus Helpdesk does, there will always be people who contrarily hate all the features that you love. It's our job to mediate that chaos.

All the helpdesk solutions out there are really just descendants of simple web-mail applications -- where some people needed to specialize for public customer service, and others needed to specialize for different needs like internal I.T. That's why the core of 4.0 feels like web-mail, and why we constantly refer to plugins when the conversation is about specialization.

That means while Cerb4 may not do everything you need in its default configuration, we're providing a platform where the tedious and redundant commonality is already done for you. We can't be all things to all people -- but if you have a strong need for something, chances are another Cerb user does too. If we provide a clean platform, and you or we provide the development resources, then any feature can be added for groups of users in a near-native, sensible, forward-compatible way. Nobody's toes are stepped on by feature creep, and nobody's disappointed by a soulless feature developed by someone who didn't need it themselves (a cardinal open-source sin).

Over time, most things will change: needs, processes, programming languages, input devices, core technologies or computing architectures. As programmers, that means our best asset is conceptual experience that's not locked into a specific implementation. As a business owner, that means your best assets are strong, mutually-beneficial partnerships with customers and vendors. Long-term, for either of us, the specific bits and bytes have the least value.

[visionit]

I'll respond to both of your specific points in detail shortly, but first I wanted to highlight something fundamental that you both are overlooking.

The core of our renewed development ethos is to share our work with the community as soon as possible.

Logically:
If we include every feature request by default, then the project turns into an incoherent mess which isn't usable for anyone.
If we only develop the features we personally need, then this ceases to be a shared project worthy of marketing to other people. Ergo, there is no reason for us to be in "stealth mode" during active development.


We're not asking, nor do we expect every function to already be included... but it's naive to expect existing users to "upgrade" to something which isn't complete. I fully understand why you've taken the decision to wipe the slate clean, but I honestly believe it's a big mistake to advertise previous versions as "an incoherent mess" or bodged together from several ideas. Yes, it had flaws... but doesn't everything? As far as I can gather, users were more than happy with previous versions! Why change a winning formula? Long-term stability is nice, but it can only be achieved if your existing client-base can remain faithful to the project. By stopping development on previous versions almost immediately, you're cornering the most pro-cerberus users by restricting the options available to them.

This long-term approach is powerful, but it's significantly different than how things were done in the past, where we'd announce a major release after a long period of silence and the result was just "how things are now". I can't imagine you'd want to return to that.

To be honest, it feels like that's happening now. Cerberus 2.7.0 (the version I'm running) was no longer being developed, moving to v3 was the now the main focus. I wasn't keen on v3, but I was left with a tough decision... Stick to 2.7.0 with limited functionality, or move to v3 which (IMO) wasn't as user friendly or support-focused as 2.7.0.

I think the problem lies in the short (or non-existent) transition period between major version changes. I respect the "transparent development" ethos and it can only be a good thing... but it's incredibly naive to expect users (especially businesses) to stay loyal to a platform which is constantly changing direction. There's no "in between"... it's either v4 or nothing. Yes, you can use previous versions... but that's essentially outdated software bottle-necked by the complete lack of updates for previous versions.

A major problem, highlighted by this thread, is that you are carrying on that outdated mindset of assuming what you see is the completely finished solution, meant to solve every problem or descend seamlessly from every prior feature. It's not. And we've been very clear in our communication, from the moment we introduced 4.0 in beta, that our goal is to rebuild Cerb4 from the ground up with the community able to provide feedback every step of the way. This isn't a box of immutable software sitting on a shelf somewhere; it's an agile, interactive project which makes daily progress.

I can't speak for anyone else, but I've never considered any software "finished", especially open-source apps like Cerberus. What I would expect is a certain amount of flexibility to users' needs... and a full re-write this far in to development I think is a huge mistake. It makes the Cerberus project look as if it's fragile and without focus. Each time I visit here (which granted, isn't as often as normal) there's a new focus to the project and it gets very difficult to keep up. The new "direction" appears sound, but so did the last one... build a platform that works... and stick with it.

What we have right now works incredibly well for the realistic, fundamental needs of everyone. We have a freshly-engineered platform that makes incorporating feedback very painless, and we have a re-energized development team that's telling you anything is possible. It's silly to underestimate the impact both of those factors will have on development going forward.

I have no doubt that the team 'will' develop a remarkable helpdesk suite, as previous versions have proven you're capable of doing so... but dropping one project in favour of a FULL RE-WRITE is just plain ludicrous. If I'd seen v4 advertised under another name (not knowing it was related to cerberus), I wouldn't give it a second-glance, despite being very well written.

The small luxuries on top of this platform, like advanced SLAs and time-tracking, are absolutely trivial to implement from a development standpoint.

It may be trivial from a dev standpoint; but it's not trivial to the end-user who relies on it to function on a day-to-day basis. My main client (council authority) will not pay for a "service" job unless I record the times for each job. One of the reasons I picked Cerberus was because it had that feature! Taking that away and dropping updates on existing versions now forces me to go elsewhere. These new changes may have been made with the best of intentions, but if dedicated enthusiasts are forced to look elsewhere (because we can't wait for a full re-write, no matter how good it is/will be), I think it's time to re-assess your strategy.

If we knew exactly what we wanted to build for such features, they would be done in a day or two from the technical-end. However, realize, at WebGroup Media we don't use time-tracking or advanced SLAs. We're very experienced software developers, but we're honest enough to say we aren't call center consultants. There are genuine experts in our community who know exactly what a feature like time-tracking should encompass. It's much smarter to ask for their feedback than for our developers to fake it (again). Agreed?

Agreed! Seek professional advice but "improve" the current system, but don't drop it completely until a better solution has been found. Vista crashes quite often, but even Microsoft don't expect you to "sit tight" and wait for an entirely new version of Vista to be written to resolve the odd glitch. The previous version, regardless of its pitfalls, was sufficiently good enough for people to use v2/v3.

[B]As a fellow business-owner, I generally don't have the patience to let open-source democracy take its course from a completely clean slate. I look for complete solutions that I can plug in to solve a problem with minimal time or cash wasted.

Likewise, and we can allow a certain amount of flexibility but a full re-write at this point in development effectively removes Cerberus from the market until it's complete.

But over the ensuing months and years, as I constantly work around the shortcomings in these prepackaged solutions, I realize that the main feature I'm missing is a responsive, transparent, personable development team who embraces change because it's inevitable. That is exactly what we offer you.

Don't get me wrong, you all do an excellent job and I'm very grateful, but I think some of the recent business decisions could prove to be the downfall of Cerberus. My concerns about the software aside, the seemingly constant changes in direction/focus means Cerberus isn't a sustainable platform and isn't a viable solution for any business looking to run a stable, long-term helpdesk.

But, realistically, some of the alternatives you have mentioned fail horribly in the long-term measures of developer transparency; including "full-attention" dialogue like this, source code availability, support, etc.

I agree there's little transparency when dealing with companies/software like Kayako, but if it's a choice between functionality/being able to run my business & speaking one-2-one with a developer, I'd pick the former. Cerberus 'had' the best of both worlds, which separated it from other projects (paid or open-source) but I believe that bonus has now gone until Cerb4 is "complete", or at least did what previous versions could do.

All the helpdesk solutions out there are really just descendants of simple web-mail applications -- where some people needed to specialize for public customer service, and others needed to specialize for different needs like internal I.T. That's why the core of 4.0 feels like web-mail, and why we constantly refer to plugins when the conversation is about specialization.

I have web-mail in the form of Horde, but I'd never use it as a basis for a helpdesk implementation. Other than displaying emails, they're like chalk and cheese.

We can't be all things to all people -- but if you have a strong need for something, chances are another Cerb user does too.

... but isn't that how it's always been? You didn't develop v2/v3 in a day, it was a gradual progression working with a community to bring more functionality as time went on. Those clients probably still require that functionality today... you've now taken that away.

If we provide a clean platform, and you or we provide the development resources, then any feature can be added for groups of users in a near-native, sensible, forward-compatible way.

Again, wasn't the first version a "clean platform" ? A re-write should make EXISTING software more efficient. If you do not re-write previous functionality, it's essentially a completely different product. It's Cerberus WebMail with a plugin to make it a helpdesk.

That just seems wrong to me. Sorry.

[jstanden]

Again, we're covering a lot of ground so I'll need a little time to draft a considerate follow-up to all your concerns. ;)

But I can tackle these key issues first:

... but isn't that how it's always been? You didn't develop v2/v3 in a day, it was a gradual progression working with a community to bring more functionality as time went on. Those clients probably still require that functionality today... you've now taken that away.That's rather melodramatic. We've only taken it away in the sense that we decided to stop splitting our official focus, and limited resources, between Cerb3 and Cerb4. Sure, we've gone "all or nothing" in that regard. Cerb3 is still open-source and the licenses don't ever expire. What's taken away?

But it's a bit premature to declare that any Cerb3 functionality has been injudiciously stripped away by us on some lost crusade for simplicity. We simply have to rebuild the lower-level (parser, ticket, worklist) features in 4.0 before we can build the higher-level stuff (time-tracking, localization).

No matter what we opt to do, or desire to do, building a sensible successor to 3.0 is going to take time. We could have been building 4.0 in secret until it inherited all the mistakes of 3.0 -- for the sheer sake of not upsetting anyone who doesn't grok iterative development -- and what good would that have done? What would be the point of rewriting and not accepting as much feedback as possible to avoid repeating past mistakes?

One could make the outdated argument that it's not professional to put a transitive product on our website, but at the same time I'd make the counter-argument that it's the most honest thing we could do. What better introduction to the project than to say, "We're not going to have big whiz-bang releases that catch you by surprise. We're going to work on what you ask for."

Perhaps your major, credible complaint is simply that we've gone to such an extreme to deprecate 3.x in the meantime. Once we started to solve problems with the 4.0 mindset, we realized that it wasn't practical to backport these solutions to 3.x. Beyond it being a waste of time, bringing the major benefits of the 4.0 code to 3.x would constitute to a total rewrite anyway. It's essentially what we're doing, except we aren't starting with all the features in place and trying to work around them while renovating -- we're re-evaluating every feature, improving it, and bringing it back in a way that makes a cohesive whole.

If I'd seen v4 advertised under another name (not knowing it was related to cerberus), I wouldn't give it a second-glance, despite being very well written.Well that's where the real irony is -- and this is important. Once we started asking new people what they used for a helpdesk previously, the overwhelming majority response is POP3/IMAP mailboxes; not a fancy competitor. When we discovered that years ago, it surprised us. (It's still true today too.)

When we're talking to people who's first introduction to Cerb is 4.0, they're generally quite happy. It's a major improvement for those people. Even for all the 2.x/3.x flaws that grated on our development team, it's still a typical case of being more critical of oneself than anyone else truly is.

But it gets a lot more interesting when the 4.0 critique is by some long-time 2.x/3.x users. And the irony is where we've inadvertently encouraged people in the past to think the wrong things are important (like manually juggled due dates, priorities, custom statuses, 1:1 routing, etc) -- and then those people look at 4.0 with 3.0 goggles on. The merits or rationale of 4.0 are thrown out wholesale simply because on the surface it looks different, and some concepts don't map over yet. These people don't stop and think objectively, they think reflexively. That's not healthy for the project.

I can only ask what the sudden hurry is? If you don't feel 4.0 is ready for you yet, then wait. It's constantly evolving, and you can constructively apply pressure in the stereotypical form of, "I'd buy/upgrade if you ________" and we could weigh that against the consensus when picking our tasks each day.

If it's just impatience, then I'm there with you. I'd love to wake up tomorrow and have the groundwork completely done. However, the rewrite has drastically improved everything that we've reimplemented -- it's been worth the risk. I wouldn't change a thing.

Again, wasn't the first version a "clean platform" ? A re-write should make EXISTING software more efficient.That's a nice platitude, but it's not realistic.

Our first version wasn't clean or a platform. We were also 6 years less experienced when we designed it.

As I've said many times in the past, Cerberus Helpdesk has succeeded far beyond our wildest expectations from those first lines of code in January 2002. None of our early code was written to support 18,000+ installations, or to be maintainable for 6 years into the future. There's a huge "WTF?" cost there every time we add a new developer to the project, or consider how to implement feedback against code that doesn't have a strong framework.

Another big problem with the Cerb 2/3 codebase, that's also invisible to you on the consumer end, was our flawed strategy of targeting the "lowest common denominator" for PHP/MySQL compatibility or server configuration.

You need to understand how pervasive that hacky 'workaround cancer' was when we'd reinvent the wheel so the application would work on 5 year old PHP/MySQL versions.

We were in a purgatory where we couldn't take advantage of modern improvements in PHP5, and we couldn't refactor out the clutter required by older versions. Can you start to see how this "tiny changes" theory is unrealistic? These inescapable flaws were so pervasive because of a flawed strategy. They couldn't be corrected without changing the strategy. And you can't inch yourself out of that trap.

I'll come back to some of your other points in a follow-up. But I still think you're being a bit knee-jerk here in assuming all the old features are gone for good, or that we're somehow naively self-destructive.

It's like you walked into a meeting that's been in session for 2 hours, and you listen to the last sentence that was said, with zero context, and you proceed to declare how everyone in the room is wrong. Slow down. Take a deep breath. Let's get some mutual understanding going here. ;)

[xps]

Hey Jeff,

first of all - i dont want you make cerberus a 'bad software' or upset you in any way. I just want to say my concerns with the new version.

Im not a developer nor do we have time here to write our own software - thats one reason why we looked for a 'finished' version. Personally i dont care that much about beeing open-source or not cause beeing open source doesnt provide that much goods for me (cause im not able to use it :) ).
I dont want so say something against open source - its just not usable for me at this time, it has no benefit over closed source that im aware of.

You meantioned it a lot of time that all can be done with plugins and that the community will bring it together - but i cant see that yet nor can I WRITE something like that. The result is that people like me will miss functionality unless someone writes it.
Like the support center which you introduced now after people requested it - come on - that looks like a 'damn we need to do something that we can call a support-center' - i dont wanna be rude and i cant say anything about what under the hood but from my stupid non developer end-user view you did already a way better job in cerb3 with it.
Same goes for the knowledge base - it really looks like a wanna-be web 2.x something thing - but lets be honest it just isnt close to look as a professional application.
We have some customers that are well known here - they would throw me out of the room if i would try to sell then this for their knowledgebase / customer support website.
They are also non technical user - they have no idea again whats driving this - they want it to look good (and work ofc) but they will more usual buy something that looks like a professional solution.
Let me compare it to kayako - just look at kayako - it looks good - it feels good and it works. Tbh i never would have switched to kayako from what i heared about their support and badly written source code. I really think that cerberus wins against kayako from the technical side of the view - im pretty sure you do a way way better job than them.
But i really think you will loose a couple of customers because the way you have taken for cerb4 - it may be only those not that much technical orientaded user that can read and praise your source - but theres a lot like me as well. I think you will gain a lot new users which is fine and brings fresh money but loosing long time customers is not that good at all.

We had a customer that wanted a helpdesk so we presented them Cerberus3 - Kayako - Auracle. Because of all the good experience we had with Cerb2 and 3 we could push them to take Cerb. That was even with 3.0 not the easiest way cause they just looked at kayako and found it looking way better. Auracle was out on interest in version 1 for them. We showed them cerb4 as well but it was denied cause it had no support center at this time. After installing it they were missing HTML mails which they wanted to use so in the end cerb3 which couldnt offer this was replaced with support trio which didnt look as nice as kayako but had all what they needed.
They also have a big call center and sell their capacity to other companys - so they need time tracking - another nogo for cerb4.

Ofc you said it a lot of times - you plan to bring this all back into cerb4 - but atm its not there and we cant rely on what may be there in the future. Like the Live-chat - we paid your for that years ago and it was always promissed to introduce it - but wheres it? Why did i spend money on somthing that i cant use?

And honestly i really cant understand why you wiped features where i would say that this was core functions from the current list.
I sit here and watch at cerb3 which i really really liked and than i watch at cerb4 and in the end i come to the conclusion that (at the moment) it is in no way a comparable product. It looks and feels like a stripped down cerb3 so you can offer something cheaper to the world and they can buy plugins as needed if the like to get the functionality of 'cerb3 plus'.
No support center, no live chat, no workstation, no time tracking, no SLA's, lots of functions missing (doesnt matter how complicated it was to use them), no 'open ticket' link......

You said:
The merits or rationale of 4.0 are thrown out wholesale simply because on the surface it looks different, and some concepts don't map over yet. These people don't stop and think objectively, they think reflexively. That's not healthy for the project.
And i must answer that OFC i have cerb3 goggles on .- what do you expect :)
And i ofc dont think objective but im not interested in whats best for the world but what best suits my needs - and if you're working with cerb3 and your used to used whats in there you check out cerb4 and just think 'wft - what have they done - wheres this, wheres that...' I hope you understand what i like to say.
Again this IS a great product for everyone entering this community from his outlook experience but it is NOT (at this time) for someone that uses cerb3 or kayako or... - i would say the simply wont change at this time. OFc cerb3 is still working but we all know what happens if a new alpha realse of something you daily use is announced - you see lots of 'when will it be released' questions cause people like your work and they like that they will get some more functions and features. You throw them in some sweets and make them hungry at the new version and then you serve than something that has 25% less features - dont expect them to be very happy :)

I really ask you to understand that i dont want to say anything bad about cerb4 in the end - i just want to explain what my thoughts are about it - i bet it is a well writen and thought over script. I dont want to be rude or upset you and your team - i bet you did a great job and cerb4 will be a great product. I just like to share my personal view about this atm because i really like cerberus and im said about the way it took at the moment. Thats the reason why i take the time to write all this long story - if i wasnt interested or bound to cerb in some way i would just move on.

Thanks
Mike

[jstanden]

Hey Mike!

I'll reply to your other comments in the afternoon (it's 3am over here!), but I wanted to quickly acknowledge your reply. :)

I really ask you to understand that i dont want to say anything bad about cerb4 in the end - i just want to explain what my thoughts are about it

You don't have to worry about upsetting us, we'd much rather hear your honest opinion. The good, bad and ugly!

...like the support center which you introduced now after people requested it - come on - that looks like a 'damn we need to do something that we can call a support-center'

You can change the themes on it, and two of them (classic_blue and classic_green) look like the original Support Center.

The knowledgebase area of the SC is going to be different than the old tree structure because of how Fetch & Retrieve works.

It's still possible we can integrate the KB and SC instances, but the new KB is meant to be slim. The KB is most useful in conjunction with Fetch & Retrieve, rather than sending customers to the KB's URL directly. It doesn't need to look fancy, it just needs to provide article content and RSS feeds.

Thats the reason why i take the time to write all this long story - if i wasnt interested or bound to cerb in some way i would just move on.

We completely respect that you guys are taking the time to post your thoughts. Your candor is also very appreciated! :)

As I mentioned above, our other option was hiding Cerb4 until everything was done. But that's seriously flawed for a number of reasons:

Cerb4 actually works fine already for many people with simpler requirements (including us at WGM).
The longer we hid Cerb4, the less opportunity everyone would have to give feedback. We didn't want to simply clone all the features we had before, since some things (project management) are unnecessary and others weren't developed coherently the first time around (time tracking). It would have been worse to add things back the way they existed before, and THEN re-evaluate them and change them again for 4.0. It made more sense to wait until we could do each feature properly.
The longer we hid Cerb4, the longer we'd be selling new copies of Cerb3 that were guaranteed to be deprecated within a couple weeks/months. If didn't make sense to sell copies of Cerb3 when we were knew we were discontinuing it. We'd rather people look at Cerb4 and say "not ready yet" (and lose the sale for the short-term) than feel like they were "bait & switched" from Cerb3->Cerb4 (and lose over the long-term). We'd rather make our intentions very clear and unambiguous.I'll reply to your other comments in a couple hours, after I've slept a little. ;)

Thanks!

[visionit]

3AM and still willing to help! Therein lies the reason why I've always stayed with Cerberus. You just don't get that level of dedication with face-less companies.

Get some kip Jeff, we'll speak to you later :)

I think Mike and I have covered most of our concerns already, but in essence, Cerberus feels like a shell of its former-self.

I believe it's lost touch with the existing communities requirements in favour of a more stream-lined, POP/IMAP driven system.

I appreciate that your limited resources greatly restrict your abilities to focus on multiple projects at any one time, but I'd strongly advise you to rethink your strategy in terms of new-app development. Cerberus v4, IMO, should be Cerberus Lite... a lightweight but extensible framework for companies that require a foundation on which to build. Cerberus 2.7.0/3.x.x (again, in my opinion) feels more like a professional version, for those who require a turnkey-style solution backed-up by 6 years of development.

V4 (even when finished) won't be anything like previous versions because of this change in direction... and I'm not saying that's a bad thing, but don't lose touch with the community that helped build previous versions. "new" users may prefer a web-mail style environment, but previous users (like myself) loved Cerberus the way it was... and the decision to drop all ties with previous versions is, I suspect, going to come back to bite you later on.

While other projects are adding more functionality and tweaking current systems, Cerberus is busy re-coding an already superb system. Don't fix what isn't broke.

Cerb4 actually works fine already for many people with simpler requirements (including us at WGM).

That really encapsulates the problem perfectly and adds weight to the argument of having a "lite" & "pro" version of Cerberus.

You can change the themes on it, and two of them (classic_blue and classic_green) look like the original Support Center.

I don't wish to sound pedantic, but what was wrong with the old support center? It was clean, quick and well-designed. It feels like you're re-inventing the wheel, but this time... it's square. Within a few hours, any rival application could be made to "look" like Cerberus, but that doesn't necessarily mean it will inherit its functionality and ease-of-use.

We'd rather people look at Cerb4 and say "not ready yet" (and lose the sale for the short-term) than feel like they were "bait & switched" from Cerb3->Cerb4 (and lose over the long-term). We'd rather make our intentions very clear and unambiguous

Very honourable intentions, but for many (myself included) it's simply not possible to wait for a new version to be finished. I am looking to upgrade because there are features which I desperately need that aren't included with 2.7.0/3.x.x. As neither of them are supported nor actively developed anymore, and v4 is completely different... you leave me with absolutely no choice but to move to another product.

For those happy to run the previous version (which I was until recently), the delay for v4 doesn't affect them in any way. Those individuals/businesses looking to improve and move forward with their existing installation, can no longer do so. You won't lose customers in the short-term... you'll lose them completely.

If, in 6 months time, Cerberus 4 returns to being a true helpdesk with all the added functionality we need, I'll have absolutely no qualms dropping Kayako and swallowing the costs involved in doing so... but I suspect many will not.

I (for one) am dreading the work involved in moving to another helpdesk system and never dreamt i need to do so just 3 years after finding Cerberus. Your decisions have a real impact on real businesses. You could build the finest helpdesk ever seen, but once you give the impression that the project lacks direction, many companies just can't afford to take that risk.

It's difficult to comprehend the amount of the work needed to migrate to another helpdesk, especially from a developers perspective... so I hope the following helps.

----
Excerpt from Migration Plan: (copied from notes sheet, so please excuse the references to other products)
1. Obtain license for "Kayako SupportSuite": £250
2. Obtain license for "KayakoMobile PocketPC Edition": £22
3. Obtain license for "Kayako Syncworks": £22
4. Obtain license for "Kayako InstAlert Pro": £11
5. 8hr maintenance window in which to backup existing platform, install and configure new platform and migrate previous data: £720
6. 2 "1 day" training sessions to introduce staff to new system (25 members per session): £1440
7. 4hr Q&A session (2 weeks later) to address any concerns & complaints from staff: £360

Total estimated cost: £2825
----

The financial impact is minimal, but it has a knock-on effect on staff who need to be pulled from regular positions to be trained on a new setup, be it Cerb4 or anything else.

I hope that puts things in to perspective. You may prefer to start-over, but I (for one) really wish you hadn't.

[jstanden]

I posted a message a couple minutes ago that's really relevant to this current discussion:
http://www.cerb4.com/forums/showthread.php?p=2175#post2175

I haven't forgotten about you guys. :) I have a big note on my desk to reply to all your key points.

[Pete Ness]

I, too, am a long time Cerberus user (3+ years, at least).

It's frustrating because each version gets "close", but then you decide to move on to something else and we're left waiting while you build the next "new big thing". MajorCRM? Cerberus Java client? LiveHelp? 2.x? 3.x? 4.x? How many of these projects are still "alive"? While you decide that the original paradigm was wrong and start things from scratch, all your existing users get to wait for you to "change the game" (when we're pretty happy with the current game) and build everything back up again so it's usable.

It seems like if you'd just pick a direction (the same direction your customers picked when they bought in), then we'd *all* be so much further along. Many years ago, there was a quote (maybe a comic) in the back of some magazine (which I can no longer remember exactly!) which had an oncoming car and a rabbit. All the rabbit had to do was pick *a* direction, any direction, and it could escape to safety. Instead, it hops back and forth until hit. I feel like if you just picked a direction and stayed going in that direction, you'd be a lot further along (and so would your customers) than you currently are.

This is what's so frustrating to all of us - we make it to a good place and can see the *next* great place, and then you decide to scrap everything and start over. 3.x was very much of a shift from 2.x - after many months of work, it did finally make its way back to workable again, but about then you decided to throw it all away again and start on 4.x. Really, it seems like everything is always in turmoil here.

Jeff, your intentions are good, and that's why a lot of us keep coming back, hoping you get it right. But years go by and I'm still waiting for some of the stuff that I asked for (and heard others ask for weekly) in the 2.4 days.

- Pete

[jstanden]

Hey Pete!

I, too, am a long time Cerberus user (3+ years, at least).

Glad to have you around!

It's frustrating because each version gets "close", but then you decide to move on to something else and we're left waiting while you build the next "new big thing". MajorCRM? Cerberus Java client? LiveHelp? 2.x? 3.x? 4.x? How many of these projects are still "alive"?

You act like freely exploring new ideas is a weakness. We don't see it that way. An aversion to change is not in our ethos.

As for project statuses, it's really the wrong question. Each of those projects has served a purpose on this progression.

Workstation (Java/Desktop): We wrote Workstation as a desktop application so we could test some workflow ideas without changing the Cerb2 web interface. We used that interface to incubate some new features, and then transplanted them to the web interface again. Afterwards, the desktop interface became redundant and was deprecated.

MajorCRM: This was an early codename for Workstation, when we were considering forking Cerberus Helpdesk into two projects. At the time we were considering going J2EE/Tapestry or PHP5. Because that was nearly two years ago, it wasn't practical to do either. When Zend announced the PHP4 end-of-life recently it made Cerb4+PHP5 a lot more reasonable.

LiveHelp: We've gone back and forth on this one. We did a version years ago in Java/Swing -- but many people (validly) complained that it didn't actually integrate with the helpdesk. This comes back to how fragmented the 2.x/3.x codebase was (with a lack of a solid framework/API). We have a waiting list with something like 1,700 people who are interested in LiveHelp; and with the 4.x codebase (and forthcoming RESTful web services API) this is something we'll be able to do much better than we could have before. We're also now competent in Java/RCP (from the PortSensor project) which is superior to Swing -- native-feeling and cross-platform -- and makes iterative product updates a lot easier, which was another big flaw in the past. Our original LiveHelp also had a ton of bloat trying to do analytics because it was developed before Google Analytics went public.

2.x: Long dead (from a development standpoint)

3.x: We're still supporting it, but it's being phased out as we transition the things that worked into 4.0.

We could have been doing this rewrite behind closed doors, but that wouldn't be very sensible. If you're really bothered by seeing 4.0 being developed out in the open, just pretend it doesn't exist until it does what you want. ;)

I feel like if you just picked a direction and stayed going in that direction, you'd be a lot further along (and so would your customers) than you currently are.

Again, that's a pretty platitude, but it's far from realistic. It's easy to say if we'd settle on one arbitrary vision, and stubbornly refuse to reconcile it with new technology or experience, that we'd have this perfectly refined system through sheer perseverance. How many things actually work out that way, especially in the software industry?

Entropy is a huge factor, especially when the original codebase was not built to scale to ~19,000+ installations. It didn't help that PHP4 was very weak on object oriented programming practices. Nor did it help our groundwork was laid 6 out years ago before we had an inkling of how popular this project would become.

3.x was very much of a shift from 2.x - after many months of work, it did finally make its way back to workable again, but about then you decided to throw it all away again and start on 4.x

We never threw anything away between 1.x and 3.6 -- and that is the problem. The ugly turmoil of 2.x to 3.x was because of old concepts not making way for new concepts. They tried to co-exist, and it was a nightmare. I learned a lesson from that, you're asking us to repeat it.

...which had an oncoming car and a rabbit
The rabbit story was cute -- but it's unfortunate if you really think our development style has anything to do with being indecisive.

Here's a counter-analogy:

Some ideas are like disposable booster rockets, in that they couldn't possibly take you the entire distance to where you need to go. They simply give you an early push in the right direction. While those boosters are burning you have a limited period of time where you can either: (a) plan for the inevitable day that you'll have to jettison them, or (b) enjoy the free ride and coast.

(a) is a stressful buzzkill, because you're thinking about the future even when the current situation feels like it could go on placidly forever. But it's the pragmatic, responsible choice.

(b) is less stressful in the short-term, because you can simply go with the flow and procrastinate dealing with the inevitable future. Unfortunately, that's human nature. We've evolved to live in the moment. But that doesn't change the fact that this choice is going to leave you stranded.

We're operating on the long-term outlook. Cerb has been around for 6 years. There's no reason to assume it won't be around for 2+ more years. The time to be thinking about 2 years from now isn't 1 year and 11 months from now.... It's now. It's something we need to be thinking about before we need it; before the booster runs out.

Jeff, your intentions are good, and that's why a lot of us keep coming back

I'm glad you feel that way. These discussions are the true soul of the project.

years go by and I'm still waiting for some of the stuff that I asked for (and heard others ask for weekly) in the 2.4 days

Have a couple examples?

[darren]

We've been a Cerberus client for a few years now, starting off with 2.x (which was quite a bit more expensive than the current pricing), and upgrading to 3.x when it came out.

It serverd us well, but the interface on 3.x was a nightmare for us, and we migrated to Kayako for one of our brands, only using Cerberus for our corporate emails which were lower in volume.

I looked at the Cerb4 demo when it came out, and was initially impressed to see that some of the functionality for mass managing tickets, bulk spam assignments etc was there. I then looked at the "support centre" and really thought WTF! Having a good client interface is essential for us as it's our only means of contact with our clients.

We carried on with Kayako, but recently I've been fairly frustrated with parts of Kayako's system. We could never get the loginshare to work properly and Kayako has no spam management whatsoever. You have to do it at the server side and then filter it out somehow. We still get lots of spam. Their monthly licensing is also a but harsh. Forget to update your key.php file and the whole system, including the client interface stops working with a big "license error" notice!
There's also no way to manage multiple brands properly from a single install. Having multiple support centres on different domains for each brand with a central admin area for all email si a real bonus of Cerberus.

So today, I took another look at Cerb4 and installed it on our own server. I spent about five hours playing with, trying various features and exploring how it could fit into our current business setup.

I decided to look at it from two angles.

As a php developer, I can really applaud the decision to start from scratch with php5 and object-orientated programming. We decided to do that with our own commercial application and it's made development a lot easier. PHP5 has matured a lot and it's OOP support brings a lot of advantages.

I like the svn install model. It's not for everyone but it does make updates very easy and faster for everyone. It would be good to have a daily build into a zip and tarball though for clients not wanting to go down the svn route - the current zip is quite a bit out of date.

As a user, I find the new worker interface (for want of a better phrase) to be cleaner and easier to use than Cerb 3.x. There's quite a lot of functionality in there, and although not everything from Cerb 3.x I've found there to be about 90% of what we actually need.

The concept of plugins was initially a little misleading. I'd always considered plugins to be something developed by third parties for use in an application, rather than parts of the main application. Maybe modules is a better way to think about it? You can then enable or disable certain modules if you need them.

Similarly, communities was also a bit misleading. Something like "client sites" or "public interfaces" might be a better way of thinking about it. We have clients who contact us so the concept of them as a community was not something I immediately recognised.

The old Cerb 3 style support centre is there, but it's a bit subtle to find. I think with a little bit of template skinning, it'll be possible to create a good client support centre similar in functionality from a client's perspective to what is available for Cerb 3.

The knowledege base is now separate to the support centre, but I like that they really are separate and per community so we can now finally have separate knowledge bases per client site/community. It would be better to be able to edit the articles from the main worker interface rather than having to have separate logins on the community kb, but I can understand the logic behind it. It would also be good to be able to share articles between communities.

Now the main thing I do like is the fetch and retrieve concept. The idea of being able to bring together search results and information via RSS feeds is an excellent idea. Why should I have to see if the client has searched the forums, read the wiki, check the blog etc. when Cerberus now does all that for you. A single unified search of whatever resources we have. Brilliant!

We've still got to make the final decision as to whether we're going to go with Cerb 4 and pay for our upgrade renewal, but given the continued cost of running a monthly Kayako, I think Cerb 4 will have the advantage.

There are some features that I think Cerb 4 should have that it doesn't yet, but it's getting there, and as Jeff has said, adding these should be simple and fast for the dev team. As long as we're getting fast updates and the obvious enthusiasm continues, I think Cerb 4 will continue to grow.

For people considering jumping over to Kayako, I'd say that from experience, the grass might be prettier, but it's certainly not greener!

If people are interested in seeing our support centre when it's finished, I'd be happy to show it off. Hopefully it will live up to my new-found interest in Cerberus.

[visionit]

There are some features that I think Cerb 4 should have that it doesn't yet, but it's getting there

I think that's the problem! Previously, we'd have been willing to wait... but we need a system that's "already there", not still being developed & debugged.

You have to do it at the server side and then filter it out somehow. We still get lots of spam. Their monthly licensing is also a but harsh. Forget to update your key.php file and the whole system, including the client interface stops working with a big "license error" notice!

I will admit, the lack of a Bayesian spam filter was a big concern, but a single installation of SEM Spam Fighter will resolve that.

We've purchased the full, non-expiring version... so licensing will never be an issue for us.

As a php developer, I can really applaud the decision to start from scratch with php5 and object-orientated programming.

Ditto. I'm not trying play-down their efforts, but I believe the decision has been taken purely to make future development easier and without adequate thought for existing clients. I appreciate that programme languages and procedures inevitably evolve, but basic requirements typically don't deviate that much.

Anyway...

Whether I agree with the change of direction or not, our existing 2.7.0 has become obsolete & v4 doesn't meet our requirements... and for that reason, I've had to migrate.

[Pete Ness]

Hi, Jeff.

I think I came across a bit too strong on the "stay the course" message. I don't think rigidity is the right answer, but you've got to admit that a significant amount of your efforts have gone towards "one step forwards, two steps back" in a lot of cases. 3.0 was an example of this - where you put a bunch of effort into something trying to "change the game", and then found that the game didn't need changing.


Have a couple examples?

1) Email confirmations having knowlededge base automatic "lookups" (i.e., the "We got your email, and we think these knowledge base articles might answer that question" in the response). We used Kayako initially, and had multiple instances of customers getting the right answer from this.

2) Ability to have a customer close a ticket themselves, without having to create an explicit account on the system.

3) Better ways to create ticket response for frequently asked questions (templates / knowledge base inserts / etc).

These are just a few off the top of my head that I saw a lot back when I was checking the forums (and posting myself).

- Pete

[networke]

I had to register, been a user since the 2.x days.

I upgraded to v3.6 and it is awesome, however v4 is a let down.

Firstly no support center is a major letdown, and then having to import tickets to v4 is a major no-no. Before, it was just running the upgrade script which handled everything perfectly.

I don't know why companies who have a hit product decide to one day start changing everything that made them successful, look at ModernBill. They've changed so much regarding their billing system, it bombed and now they've been bought out and people are leaving in droves.

Requiring v4 to support all these additional extensions (e.g. mailparse) is just going to make it harder for more people to install and support your products unless there is a HUGE speed improvement, which I don't see in v4.

Guys, please keep supporting v3.6. We don't want to migrate over to Kayako, because unlike you, they've STUCK to what made their product unique.

Whenever you change something, you just alienate all your users. A better way is to introduce a yearly subscription fee, so you can make more money off your existing base, they'll understand and will keep supporting you.

If you want to tap into a different market, then release a SECOND product, like Cerberus Enterprise... but please NEVER EVER change the core product and kill off your old users.

I will stay on 3.6 until there is a new release that lets me upgrade WITHOUT having to IMPORT all my tickets again. I'd be happy to pay for the upgrades.

[Hildy]

I had to register, been a user since the 2.x days.

Thanks, we really appreciate the support of our long term community members.


I upgraded to v3.6 and it is awesome, however v4 is a let down.

Firstly no support center is a major letdown,

There is a Support Center Community Tool, and it was in fact just recently revamped to integrate the KB and make it nicer. There's also the whole Web-API, where you can build a custom Support Center to exactly fit your needs. ( http://wiki.cerberusdemo.com/index.php/Web-API )


and then having to import tickets to v4 is a major no-no. Before, it was just running the upgrade script which handled everything perfectly.

The reason tickets have to be imported is because we were starting 4.0 with a blank slate, including the database. This is covered in more detail below, but the db specifically was cluttered with a lot of useless tables, data, etc, and a clean db is a much nicer place to start. (You mention liking the updater in 3.x; the 4.0 updater is fully automated and will know when files have changed and it needs to run.)


I don't know why companies who have a hit product decide to one day start changing everything that made them successful, look at ModernBill. They've changed so much regarding their billing system, it bombed and now they've been bought out and people are leaving in droves.

Requiring v4 to support all these additional extensions (e.g. mailparse) is just going to make it harder for more people to install and support your products unless there is a HUGE speed improvement, which I don't see in v4.

We didn't just decide one day to change what made us successful. In fact, we decided (and not in one day) to start over from scratch *because* of what made us successful: a passion for handling helpdesk email problems, and doing it well. With 4.0, we've focussed on that core responsibility, and we've made vast improvements over 3.x.

The 3.x codebase (and db, as referenced above) suffered from over five years of incremental development on a product that wasn't designed from the start to be capable of what was now being asked of it. Code maintenance, bug fixing, and new features / improvements all suffered greatly because of the state of the 3.x codebase, and there was no way to rescue that.

4.x was born of a frustration with the limitations of 3.x, and it's true genius is in the parts that most people will never see. Built on a pluggable base we call Devblocks, using a Model-Controller-Viewer paradigm, 4.x is incredibly easy to expand upon, to improve, to fix the occasional bug, etc. As a result, development response can be much more rapid and in tune with the community and their needs.

Because we were developing 4.0 from scratch, we have been able to take advantage of the many improvments to PHP and the third-party libraries that we use. This gives us much better error handling, faster email handling, and even little goodies like the graphs in the reports section.


Guys, please keep supporting v3.6. We don't want to migrate over to Kayako, because unlike you, they've STUCK to what made their product unique.

Honestly, there's nothing we can do to prevent you from migrating to Kayako if you think that's a better fit for you. Obviously, we'd rather you didn't, and we do feel we have a superior product, but the choice is ultimately up to you. You will note that, even though we're not officially selling 3.x anymore, you can still request a 3.x license instead of a 4.x license.


Whenever you change something, you just alienate all your users. A better way is to introduce a yearly subscription fee, so you can make more money off your existing base, they'll understand and will keep supporting you.

Not all change is necessarily good, but not all change is necessarily bad, either. Things like "peek" are a *huge* hit across all types of users (new to 4.0, 2.x upgrades, 3.x upgrades). Pile Sort has revolutionized mail sorting, and turned the chore of sifting through thousands of emails into a simple few-minute task ( http://cerberusweb.com/tour/screencasts ). Fetch & Retrieve brings together all of your online resources for your helpdesk personnel: blogs, wikis, FAQs, forums, and more. Rest assured that we haven't made a single change out of anything other than an desire to make Cerberus better at what it excels at -- handling your support problems.

Yearly subscription fees are something we don't really want to do. Once you've bought Cerberus, you own it, and we don't feel you should have to pay us over and over again for the same product. If you buy 4.0 today, you'll get any and all 4.0 mainline upgrades for the life of the product, and any 5.x upgrades free within the first year.


If you want to tap into a different market, then release a SECOND product, like Cerberus Enterprise... but please NEVER EVER change the core product and kill off your old users.

Oh, we definitely are not a one-horse show... have you seen PortSensor? ( http://portsensor.com/ ) There are also other ideas percolating in the dev cauldron; Usermeet to bring the community and the agile development team together to collaborate on product dev; LiveHelp as a customer<->helpdesk chat application, integrated with Cerberus; project management, etc.

The key thing here that I disagree with your assessment of is that we changed the core product... Sure, it's a different codebase, but it still fulfills its core mission, and does so in a much better way.


I will stay on 3.6 until there is a new release that lets me upgrade WITHOUT having to IMPORT all my tickets again. I'd be happy to pay for the upgrades.

If it's the upgrade itself that is the issue, we do offer a data migration service. You could send an email to sales@cerberusweb.com about that.