…since sliced bread

Wikis, as a rule, suck.

I say this as someone who spent the last year and a half working at what is generally considered to be a “wiki company“. I hated wikis before I went to Jot, and now having seen the strange cult of wiki even closer (Jot’s competition), I haven’t been dissuaded from my view. In return for a single interaction win, WikiWords, wikis throw away most of what is obvious and basic usability design. Had Jot not been committed to doing WYSIWYG in a wiki right, I would never have gone to work there. Heck, the one thing that 37signals got right about writeboard is that users don’t want a wiki. They totally missed the point that end-user simplicity sometimes comes down to hard engineering, but that’s for another day.

Users pick wikis (and tools like sharepoint) because they want to share the stuff they’re writing. Not because they think that wikis are cool or somehow superior to everything that came before and/or after. Sounds obvious, huh?

WikiWords as the potential basis for an interesting collaboration tool: good. Wikis as a religious creed about some sort of artificial “simplicity” related to one-off markup languages: very bad.

Today Jot quietly launched a new major version, 2.6. This version includes a brand-new editor component implemented as a Dojo widget. That it’s Dojo-based isn’t something to be proud of in itself, but the new editor provides fixes to some of the most serious issues that plague WYSIWYG on the web today. Instead of taking you to an “editing” interface for a page, the new system uses Ajax-based editing to “upgrade” the page in-place so that editing is as fluid and natural as reading.

No more forgetting what it was you were going to change as you wait for the editing interface to load. Just change it.

Perhaps more importantly, though, the new editor component does everything it possibly can to avoid data loss. Browsers do a pretty good job of keeping data that’s been filled out in a form around in case you accidentally hit a link or otherwise go to a page and want to come back to where you were. WYSIWYG editors, by comparison, totally flub this since they aren’t native components that the browser knows to save state for. Jot’s new editor fixes this with some sleight of hand, and the hope is that you won’t even notice it. That it Just Works (TM) is its best feature.

The new editor works on IE, Firefox, Safari, and Opera 9 (modulo some Opera bugs), doesn’t break the native undo stack, instantiates much faster than other WYSIWYG controls, and doesn’t impose the lame “dueling scrollbars” scenario on users. Instead of having independent scroll bars for editor content and for the page, Jot’s editor supports a “fixed toolbar” system that allows it to grow in height with the content, but never loose the editing controls at the top of the page when in editing mode. All of this has been made available as Open Source software in Dojo 0.3.0 via the dojo.widget.Editor2 component.

While the Jot editor is a thing of technical beauty, hopefully users will never think about it. Having been working with it for months, I can say that while I don’t think about using the new version, I certainly become annoyed when I run across a wiki that’s not using it. The experience isn’t as fluid…everything else just feels clunky by comparison. The new editor is the system we all envisioned a year ago when Paul first started evaluating how to improve the WYSIWYG experience. By doing our homework about what makes a great experience and by staying true to the mantras of “never break undo” and “WYSIWYG, not semantics”, Jot now sports an editor component that is second to none.

Of course, the Jot team should also be tremendously proud of the Tracker v2 (screencast) release that is now integrated into Salesforce. I’m not sure when it’ll roll out to current Tracker users, but having a look at the screencast should make users of Tracker v1 salivate for the new version. For most interactions, it’s nearly indistinguishable from Excel.

From here on out, I’m just an observer to the stuff Jot is building, but the new Tracker and Wiki releases help reassure me that our time was well spent. These products get out of the users way and help them get their work done or help them to express themselves better, but do it with the advantages of the web built in from the ground up. These aren’t products that can’t think outside the monopoly, these are tools that make people’s lives easier…because that’s what they were designed and engineered to do.

I’m proud to have played a small part in building them.


  1. Posted May 25, 2006 at 8:09 am | Permalink

    Amen, brother. I feel the same way: nice idea, lousy religion.

    I will have to try out this WYSIWYG editor of which you speak. My vision of edit-in-place is the same as yours, and I’ve done some prototyping of it (but as native apps using WebKit). I know Safari / WebKit have some bugs that make HTML editing difficult (and I know other browsers have other bugs) so I’m impressed that you (and other folks at Jot) have gotten it to work cross-platform. Congratulations!

  2. Posted May 25, 2006 at 11:32 am | Permalink

    Haha. Amen. My general dislike of wikis is what dissuaded me from trying harder to get a job at Jot. Well, that and the fact that they are in Palo Alto.

    I’m very happy that Jot Live actually happened, though, and I’m proud to have played some small part in realizing that vision. The Tracker v2 screencast is mindblowing as well. I wish Jot the best of luck, and look forward to see what other envelope-pushing things get produced!

  3. Posted May 26, 2006 at 12:35 pm | Permalink

    Have you seen WikiWyg (http://www.wikiwyg.net/)? It seems very similar to what you’ve described so far, although I haven’t tried out the new wysiwyg from Dojo.

  4. Posted May 26, 2006 at 1:04 pm | Permalink

    Hi Michael,

    So when I worked at Jot, SocialText’s Wikiwyg was one of “the competition”, but it had (and still has) several fatal flaws. The first is that when you got into wiki-markup mode and then go back to HTML, you lose your changes. Of course, there’s no notice of this when you go *to* wikimarkup mode and it doesn’t give you a chance to *not* lose your data if you are somehow duped into clicking those inviting mode buttons at the top. Data loss == a crappy experience.

    The biggest problem, though, is that Wikiwyg seems to assume that having access to Wiki markup is somehow a *good* thing. It spends amazing amounts of effort trying to convert HTML to wiki markup (but not the other way around, hence the data loss). It is, in essence, a wiki markup editor with some HTML bits hanging off the edge.

    Jot/Dojo’s editor by contrast is a WYSIWYG editor that happens to interface with a back-end which handles conversion/translation to and from the wiki markup. Which is where it should be happening in the first place. Having to keep a JavaScript version of your perl-implemented one-off markup language in sync with all of the server conversion idioms is a recipe for bad corner-case behavior.

    I remember reading a press release at some point that said that Wikipedia was going to be integrating Wikiwyg, which I suppose makes some sense if the point is to have a front end that is compatible with the wiki religion and perhaps Socialtext is shipping it to their customers, but I can’t imagine that it’s a good experience.

    If users wanted wiki markup, they would have picked the wiki markup editing mode in the first place. Our experience showed that users just care about getting their content into the wiki.


  5. Bob Haugen
    Posted May 29, 2006 at 5:21 am | Permalink

    Hey Alex, is it just wiki markup that you don’t like? Or is it also the whole freeform collaborative editing and page linking?

    (Ward Cunningham said somewhere that he wished he had a good WYSIWYG editor in the first place…)


  6. Posted May 29, 2006 at 8:47 am | Permalink

    Hey Bob!

    It’s mostly wikimarkup that I find offensive, but the total lack of structure in most wikis is also tremendously problematic. That Jot lets you build things that are as structured as you need them to be (but no more rigid than that) was what really closed the deal for me. I’m convinced that a semi-structured content repository like Jot is the future. Keeping the barrier to entry for adding content low is a wonderful thing, and systems that build in saftey nets to solve some of the problems that arise from that strike me as a “good deal” from an interaction perspective.