Opening shots in a wiki war?

Briefly - no, I don't think so.

It all started with Matt Marshall's article coining the term 'wiki war', which got picked up in various places, Graham defended Jot, Ross went on the offensive, Graham again played D, more follow up and finally Oliver came through with a well balanced summary (although I'm not sure yes, his title of "Murder, Death, Kill" is intentionally ironic). Are you still with me? Aren't blog threads fun?

A Little Aside
Before I get to my main point about openness, I would like to make one small aside. Despite how it looks (and sounds) from reading Jot and SocialText's blogs - there are more than 2 companies which sell wikis. I know of at least 5 or 6 in fact.

I say this not as the petulant kid in the corner  jumping up and down screaming "Pick me! Pick me!" but as a wiki vendor with more than 300 enterprise customers.

Customers switch vendors regularly, one customer moving from one vendor to another does not maketh a war. Once wikis are as prevalent as email in organisations, one could conceive of a war. Right now, we're all battling more to get wikis into organisations than stealing away customers. Grow the pie, don't fight over the crumbs.

Apart from this recent scuffle, the wiki industry (can we call it that?) has been very cordial and friendly. Insightful blog posts, frank discussions and generally open, friendly business practices. I hope these trends continue.

A Question Of Openness
Perhaps the most interesting actual discussion in all these threads is about the various degrees of openness involved.

Ross is trying to cast Jot as the "bad proprietary vendor":

Proprietary Platform vs. Open Source Application 
Zero-sum thinking is lazy, and quite frankly, old-school.

Here I have to agree with Graham:

I’d like to suggest instead that it’s “old school” to focus exclusively on the openness of the code when the openness of the data is at least as important.

We compete with Open Source solutions every single day, with both our products. We're used to it. Customers don't buy "because it's Open Source". Source access is only one of a number of value points customers consider when purchasing (along with price, service, features and a host of others).

(On a side note - Ross - your Kwiki page is the same as your Products page so it's impossible to tell how open source SocialText really is :))

As a customer, I'd use an application with open access to data over an application with only open source.

Data Opens Doors
Take mail clients as an example. I don't feel locked into a mail client which uses mbox format for storage - no matter how proprietary it is. I know I can take my mail boxes and move to another client that supports mbox. I don't use Outlook, not because it's proprietary, but because I can't get my mails out of the damned PST file if I need to.

We build products which support this vision. I think I'm right in saying we have more ways to interact with our wiki than any other. Confluence supports HTTP, SMTP, REST, RSS, full XML-RPC and SOAP interfaces, as well you can deploy Java code into the application, use the plugin system to deploy extensions, talk to the database directly via a published data model or use the full XML import/export facility. That's 10 full data access methods for those out there keeping count.

(For reference, all Atlassian commercial licenses come with the full source code to the application under a developer source license.)

Graham says that "open data is at least as important as open code", I'd say it's more important.

Open Source Company Generations
One of my pet theories (for those at Javapolis at Belgium who heard me propound this theory ad nauseum, you may read on) is that we're in the third generation of software companies, when talking about Open Source.

Generation 1: Proprietary Fear
The first generation of companies are traditional proprietary software companies - eg Microsoft. Their business models were forged in the last two decades, their source is proprietary, their components are proprietary, they've been around a long time and they're afraid. They think Open Source is going to put them out of business because everything is "free", hence noone is going to buy their software.

Pretty much every software company was in this boat when Open Source appeared over the last 10 years, and quite a lot still are.

Summary: Open Source = fear & doom.

Generation 2: Open Source Tunnel Vision
The obvious knee jerk reaction to this? All software will be Open Source. Open Source will completely dominate and eliminate proprietary software. Therefore we're going to make everything open source and make money off services, installation support etc. These are your classic Red Hat, MySQL or JBoss type companies.

This was a very common 'late 90's' phenomenon, and you still see it happening.

Remember - Open Source is very good at infrastructure software - stuff where there's a protocol to implement, or a server to connect to. Those software components are all about function over form, which developers love. Open Source is very bad at applications in general (possibly with the exceptions of IM clients or IDEs, but those don't really count). A good application has to be usable, form over function.

Summary: Open Source = the one true way.

Generation 3: Living In An Open Source World
What we've seen over the past 2 years though is the re-emergence of profitable proprietary software companies. How can this be? Hasn't Open Source has doomed all proprietary software companies to fail? Not at all.

In almost all cases, Open Source has enabled these companies to exist. It provides a lower cost of software development, smaller software teams and thus better, cheaper software products. These companies have grown up with Open Source both as a competitor and an aid. They respect it, but don't fear it.

(As an interesting general rule, all the examples of generation 3 companies I can think of have only 10-20 employees - like Jot (10?), SocialText (10) or Atlassian (17).)

They realise on the one hand that they must provide more value than an Open Source solution to survive, but on the other they realise the value of using and contributing to Open Source components that reduce their overall cost of development.

Summary: Open Source = pragmatic software tool

The point of the theory? When looking at companies I no longer classify them as "Open Source" or "not", I look at how much they get Open source as a useful tool to their business and how much they respect Open Source as a competitor.

Incidentally it just occurred to me that the whole concept of 'openness' in this post is exactly what wikis are about - making the knowledge of a team, department or organisation open. Editable by anyone. Can you get any more open than that? Can we call a wiki an "Open Source website"?

Comments

[Ludovic Dubost]

Hi Mike ! Happy new year ! It was nice to meet you at JavaPolis in December !

Interesting to see the US venture capital back companies fight for "visible" supremacy, when in reality the leader of open-source Wikis is probably TWiki and or professional Wikis is probably your fine Wiki !

Now to your subject: "openness"

I don't so much agree that "Open Source is very bad at applications in general". You might say that currently Open-source hasn't really delivered in this area, but from my point of view it is not because of open-source, but because open-source business are not yet there, and most open-source applications don't have businesses to back them. This will happen, and I don't think there will be much difference in terms of application quality once the businesses are in place.

Concerning the importance of data vs the code. Lock-in is both about data and code. Easier data compatibility reduces the switch cost, but unless the application is 100% commoditized switching has always a high cost. I'm surprised you haven't talked about the Wiki Syntax as a data issue. It is very nice to have all these APIs you are talking about the retrieve the data, but with no Wiki Syntax Standard, there is an openness issue. One of the reasons I decided to use Radeox as the wiki syntax engine, was to be more "open" as there was SnipSnap and Confluence using the same engine. But even with that, given the ease to change the syntax rules in the radeox engine (Confluence has many of these, as I could see) and the amount of "proprietary" macros in all the engines, currently I don't think switching content from one wiki to another is that easy. We'll see when I get one of your french customers :)

In any case, for me the two big reasons for choosing open-source were:

- Basic wikis are commodities as there are tons of wiki engines on the market.
- Wikis can be a good platform to integrate other content (with all the macros and extensions) and I believe the platform should be highly open.
- Wikis come from open-source. I learned everything I know about Wikis from open-source. Having an open-source strategy for a Wiki offering is the least I could do.
- As you say in your last sentence, Wikis are about openness ! Wikis are open-source web sites !

I had the same thought about software companies today being smaller. Indeed open-source or not, making good use of the existing open-source software allows and forces your company to be more efficient. Companies do have new options implement software and this is a whole new challenge for software companies.

[Oliver Thylmann]

Yes, that title was intentionally ironic :)

Great write up by the way.

[Jonathan Nolen]

Good post. I completely agree that we should be as concerned open data as we are as open sourcecode. And I liked the "open source website" definition of the wiki -- I've used that myself in discussions.

Just out of curiousity, besides JotSpot, SocialText and Confluence, who are the other commerical wiki vendors you mention?

[Mike Cannon-Brookes]

Well, Ludovic over at xwiki.com is one other commercial vendor I know of (Open source software, pay for services). EditMe.com is another one.

On the topic of openness, you may also want to see Jonathan's excellent (unprompted) write up of why we're an Open Company:

http://www.jnolen.org/blog/2005/01/stand_down_wiki.html

Thanks mate, I owe you a beer next time I'm in SF.

One of our customers in London reading the post, who is a lot better at being succinct than I am, summarised it as follows:

"Most open source monkeys see open source as 'low barrier to entry' - real openness lies in 'low barrier to exit'."

Perfectly put.

Mike