Lightning drive-by

We’re looking to get a Lightning 0.8 out around the end of this month, in order to start to get user feedback on it, and shift the development model to one that’s a little more conservative in terms of architectural overhauls. We’re also running into decisions with our user interface and behaviours that we would like to make with the benefit of some more actual usage data, and want that coverage as well as exposure to the various kinds of real-world data and operations that will help us robustificate the lower-level components.

We wanted to reach this point with Lightning quite some time ago, or at least I did. I’ve been misestimating software schedules for more than a decade now, and I’m really starting to get good at it.

For one thing, the cost of doing the infrastructural overhauls for a more server-friendly calendar core was greater than we thought, especially as we were trying to keep Sunbird breathing as well as bring up a different calendaring model in Lightning. I think that cost was worth it — the architecture we have in calendar/base is one that we’re pretty proud of, and it’s the result of having to make changes to accomodate actual application code that wants to deal with exceptions, and recurrence, and the ICS standard, and present a decent user experience through all of it.

The usual problems with competition for developers’ time, unpleasant dependency cascades, and the occasional bit of feature creep also caused us to slip. Nothing novel there, we just get to take that pain in public more than some other shops. (Perhaps not as public as we could have been; wiki updates and blog postings have been fewer and farther between than would have been what the kids call “ideal”.)

I’ll post in the future a little more about the details of Lightning 0.8, how we’re framing our goals, and what we expect it to look like. I’m pretty excited about getting it out for the world to see, and even to hear the inevitable criticisms of the choices we’ve made so far. Bring it on!

11 comments to “Lightning drive-by”

  1. entered 22 June 2005 @ 6:34 pm

    How long until we get nightly builds of either lightning or thunderbird with lightning integrated?

    IMHO that’s the point where we are all going to hop on and start creating a billion+1 bugs about UI improvements and start contributing code.

    Even with code available, it’s vaporware until it has routine builds. That’s when it becomes a community effort.

  2. entered 22 June 2005 @ 7:00 pm

    I agree with Robert. Nightly builds of Thunderbird + Lightning is where a lot of good attention will be given to the project. Heck, Slashdot has ran stories on Firefox and Thunderbird from pratically day one about big upgrades to early releases (Thunderbird 0.1 for example).

    I am so excited about Lightning, but I can’t help out without a build to test!

  3. entered 22 June 2005 @ 7:21 pm

    No, it’s not vapourware. We’ve had a few contributed builds up on and posted about relatively quietly, but the software exists, and is available if you want to verify what it does and participate; to group it in with software that is announced but never made available to anyone is misleading, and insulting. As it is indeed to describe it as not being a “community effort”; that it’s not well-packaged for easy user installation doesn’t diminish the contributions of the calendar community (developers, testers, feedback-providers, other helpers) in any way, even in the face of your offensive rhetoric.

    Nightlies will be available as soon as the build team has the cycles to set up the infrastructure for it. I don’t know how to make that happen faster, or how to determine the date on which it will happen. If I did, I’d have done or said so.

    You don’t need to tell me that we will get more user feedback once we have user-installable releases available for wide distribution. You might note that I said almost exactly that in my post.


  4. entered 23 June 2005 @ 1:22 am

    Great news, Mike!

    Is Lightning going to be distributed as a TB extension? Or is it tightly integrated?

  5. entered 23 June 2005 @ 1:31 am

    ah, nm. i found the xpis on, which answered my question :)

  6. entered 23 June 2005 @ 3:09 am

    I would love to find a pointer to server software able to talk with lightning, could you provide such pointers here ?

  7. entered 23 June 2005 @ 5:55 am

    “IMHO that’s the point where we are all going to hop on and start creating a billion+1 bugs about UI improvements”

    And that’s a good thing, is it?

    “and start contributing code”

    I would imagine the people that can and will contribute code are likely to be people that can build the app themselves (or at least people that will find the builds that are already available)

    The world and their dog complaining that the software is full of bugs and lacking lots of features they want is something that can happen based on 0.8 – that, as Mike said in the post, is the point. Putting out nightly binaries just means you’ll get the same comments on every build until things get fixed (which is mostly what happens with Firefox, and it’s well past “1.0″…)

  8. orrin
    entered 23 June 2005 @ 10:22 am

    i can’t wait to try it out!

  9. entered 23 June 2005 @ 12:14 pm

    Are lastest Lightning sources in Mozilla CVS?

  10. Daniel Beckham
    entered 23 June 2005 @ 12:46 pm

    Until Lightening supports Palm and PDA integration, it doesn’t compete with iCal or the myriad of software already available on the PC for cheap or free. To me, it’s interesting, but worthless.

    I’d love to use it, but why bother when I end up having to maintain two different calandar/contact applications, one for day to day use and one to keep my PDA updated?

  11. entered 23 June 2005 @ 1:48 pm

    The Lightning source has always been in the tree.

    If Lightning doesn’t do what you need it to do, you should not use it — and you certainly shouldn’t use it if it’s worthless. In my interview with sipaq and asa I talked about the status and plan for sync, so I won’t repeat myself here.