MozillaZine

Brendan Eich Discusses Future Roadmap Plans

Michael Lefevre wrote:

On 2004-04-22, Gervase Markham <gerv@mozilla.org> wrote:

[snip]

  • Need to split the roadmap into two parts — soon and future — and get the "soon" part out there

Gerv didn't fix the meeting summary per mitchell's mail saying, in reply to that "Need to split" item: "Not quite. I said that the roadmap had already been organized this way, we had the place to put the 1.7 info earlier, we just didn't do it."

This is a tempest in a teapot, but you raise some good points. I don't have all the answers, or there would be a new roadmap up already.

Below I give my opinions, but much of what I write is tentative, and we haven't gained agreement among drivers and others on the mundane stuff sketched below. The more forward-looking technical ideas I'll leave for a later post. The timeframe to get our story straight for the next year or so is the next two months, maximum.

I hope this doesn't mean "put together a roadmap for the next 2 months which everyone already knows about, and leave the future part for another indeterminate of time, and publish it when it's no longer in the future".

No, there's not going to be a futuristic roadmap; it would be useless, akin to blogging (ahem) about the future, when food comes in pills and cars fold into suitcases. Mozilla is about the next six months to a year, and that's one roadmap, not two. But it takes time to get the right plan together.

We still haven't retired SeaMonkey to a branch, which was last year's plan. It wasn't the right plan, exactly, for that timeframe, although we didn't know that until after it was "the plan". It was the right plan in that Firefox is the future of the browser at mozilla.org, and for many partners, some of whom can get us significant distribution (fingers crossed).

The original roadmap update last April proposed to retire the suite in one quarter-year milestone, which was not enough time. Ironically, when hyatt and I wrote that update, we did not know that by July AOL would be dropping the Netscape effort, leaving people out of work or looking for jobs. All we knew was that we were facing cutbacks of some kind, and that between the suite and the new apps, the better choice for any future mozilla.org was the new apps.

Nor did we know that companies looking to upgrade from Netscape 4.x would want the suite soon, last year and into this year. Such companies generally contacted Netscape/AOL personell, not mozilla.org staff, because mozilla.org was not its own legal entity, and we didn't get in the loop until July.

So last year's roadmap, with a few adjustments, is still up, and still not fully implemented, but we will be retiring the suite to a branch at some point in the next year, in my judgment. After that point, we will be making incompatible changes on the road to a Mozilla 2.0. There will be a lot of 1.x compatibility, some of it optionally bundled with apps, but not enough to carry the suite as well as the new apps.

The "soon" part was needed 6 months ago.

The Mozilla Foundation was just getting staff rehired or transitioned then, and Firefox was facing renaming (again) and trying to get a 1.0 schedule.

A top-down "general orders" roadmap would not have helped those efforts — we already had one that said "let's work on the new apps". People who for whatever reason prefer SeaMonkey are not going to help, and were not, and did not. What's more, SeaMonkey hackers may not do what the Netscape 4.x customers want — the hackers may wish to innovate in ways that break "legacy" expectations.

So I think we had the "soon" part of the roadmap you say you wanted, then. What more did you want?

The future part is needed soon.

See http://www.mozilla.org/events/dev-day-feb-2004/mozilla-futures/, http://www.mozillazine.org/articles/article4584.html, and followups.

You've probably seen them, but without a roadmap, are they useless? I think not, although some ideas (Eclipse-based IDE for XUL) went over like lead balloons. Saying "Eclipse" gets some interest, but no one to hack; a XUL IDE arguably would be better done on the Mozilla platform, but we lack volunteers who can pull it off. This is just one example.

There is a lot going on behind the scenes, not all of it worth talking about yet, and certainly not ready to road-map. But it might help to know that we are working hard on requirements and plans for the future.

None of this should disrupt 1.7 or 1.8a, although 1.8 post-alpha, and future milestones, may get a new schedule. More in a bit.

Getting the soon part soon and the future part in the future is useless — surely we're supposed to be following the roadmap, not mapping the route already taken.

Right.

Splitting the roadmap into parts has been mentioned several times in the past 6 months or so. The meeting minutes from last October 3rd, when Brendan was just about to send out a draft...

The revision happened without a split. The table-of-contents was added, and the bit about dropping the suite was updated to talk about paying customers keeping it alive, but only until the payments stop.

Volunteers may wish to carry the suite on a cvs trunk, based on the API push to 2.0, after cvs.mozilla.org puts it on a branch. I don't see why it should be on the trunk at cvs.mozilla.org, however. We won't be closing the tree if the suite breaks, and we will break it sooner or later, intentionally or not. We try not to host dead code. Volunteers could keep it going at sourceforge, tracking API changes and so on.

Enough about the suite. The next roadmap update will posit successful, continuously improving apps (Firefox 1.0, Thunderbird 1.0, others to be announced), and will address platform issues as much as application architecture. It will shake things up, which will require putting old code on a branch for true sustaining engineering.

Mozilla 2.0 is the platform push (chewing bubblegum), with killer apps (walking) something we continue to do well, as well. At least that's my view.

I'll blog now, I promise.

Got a response? TalkBack!