Open-Xchange: an email system only for the brave

Take a long look at Open-Xchange website. Looks pretty slick, doesn’t it?

I encounter many people who are shopping for something that isn’t a full copy of Microsoft Exchange (although in my opinion they should be examining the latest versions of Exchange, even if they have scars from earlier ones). The market in email packages is changing shape: there are mobile devices; there’s roving access; there’s group working; and for me as an old greybeard, there’s the odd spectacle of a younger generation who’ve grown up with paltry and cut-down ISP-hosted email boxes that you pay for by the megabyte.

I think OX has made a world-class rod for its own back

The truth is, these early proprietary email products for business – Exchange/Outlook, Lotus Domino, GroupWise and so on – did a good deal more than any ISP-supplied system ever could, given their privileged position inside the company network. Internal email installations can be the most organised, structured and collaborative bits of software that a regular business user has contact with. If you suspect you have users like this, the test is simple: threaten to take away, or even slightly dilute, their feature list, and wait for the screaming. You’ll know soon enough.

Later offerings, which arrived along with widespread web adoption and the emergence of ISPs and web hosting, look curiously limited by contrast. It took pretty revolutionary moves by the biggest players, most notably Microsoft and Google, for the rest of the hosted-email marketplace to relax their draconian space and usage limits to something more in line with small-business needs. Even then, the classic POP3/IMAP mailbox was a solitary beast.

It was with this dichotomy in mind that I was so pleased by the burst of PR work that brought Open-Xchange (OX) to my attention, first of all via several encounters at shows, then latterly with a face-to-face meeting with the boss. Sixty million users, he said; class-leading mobile device access and group document editing, he said; rock-solid dev team back in Germany. Plus, an online demo (not working during our meeting, but hey) freely usable for a test drive, and free to download for Linux installers.

I took the demo and, based on what I’d seen there, recommended that a client seriously consider OX, too. Then I popped back home, intending to try the downloadable demo VMware appliance, pleased with the slick and small-business-friendly nature of the whole deliverable. Regulars will know that Linux isn’t my forte, but based on other recent experiments, I figured that its product development and my naivety might be just about at the crossover point. Oh boy, if I’ve ever been more wrong, I can’t remember when.

Nothing worked. I suspected something was wrong with the instructions, so I mailed the contact I’d been given, who replied that I shouldn’t try to install OX on a Windows server. I should have got out then, but I decided to give a proper Linux sysadmin a straight crack at it. He found two or three basic syntax errors in the setup instructions, sorted out the non-startup of X Window in the demo virtual appliance, then left me to it.

I didn’t get much further, but my client did. Intrigued by the demo (and not much more clued up about Linux than I am), he called in the big guns, rather than an occasional sysadmin: not just someone who talks Linux, but someone who owns and is the principal architect of his own Linux distro, writes regularly for The Register and has been known to contribute to PC Pro from time to time, namely Liam Proven.

Had matters proceeded according to plan, Liam would have typed in a couple of hundred characters’ worth of shell commands, whipped up a few package update/download combinations and my client would have had a running, internal, VM-based groupware and email solution. Instead, Liam and the OX team have been mailing each other for months, with bug checks and revisions and clarifications to the English OX setup instructions that showed him, and me, that this is in no way a finished, shippable product.

What’s worse is that the OX marketing and support teams treat every question as if it has come from someone like me, and so send back idiotic links to FAQs and RFCs that aren’t relevant to the questions that have been asked. Then, when that didn’t work (Liam being a very determined chap), they said that without a support licence, even the demonstration couldn’t be expected to work.

That left me rather speechless. What’s the point of providing a package, comprised of both your own code and your instructions for using it, that simply won’t work as presented, and then attempting to charge people for telling them that?

Overall, I think OX has made a world-class rod for its own back. Full marks for attracting the attention of hard-pressed business email admins with the product name, online demo and extravagant claims, but absolutely nul point for the subsequent execution and customer relations when the gaping holes in documentation, version control and ease of use become clear.

To take advantage of a market gap created by difficult setups and harrowing lifecycle costs, you actually need to deliver something that’s easier for people to use than your competitors’ products. I’d directly quote Liam’s opinion, given his undoubted expertise in this particular field, but I suspect our legal team wouldn’t let it through…