This site is an archive; learn more about 8 years of OpenHatch.

Pivotal Tracker

by Asheesh September 14th, 2009

A few days ago, my friend Terence introduced me to Pivotal Tracker. It’s a web service for planning a software project.

It has a few fantastic aspects:

  • It is beautiful and a pleasure to use.
  • It helps remind us to review each other’s work.
  • It is real-time; if I wonder what someone is working on, I can see that immediately without even leaving my own task list. What a stress relief!
  • It helps us see how fast we are moving through our project. However fast our “project velocity” is, Pivotal Tracker adjusts later weeks so that our expectations are always realistic.

Even though we are allowed to use it for free, we give up some control to use it.

The technology is built by Pivotal Labs, a software shop with a lot of experience in “Agile” development. Pivotal Tracker puts together what they have learned; everything from the documentation to the tool itself serve as education for us in Agile strategies. Besides the educational benefits, it boosts our efficiency. Before we found Pivotal Tracker, we were using a shared spreadsheet that contained our goals, or “user stories.” We were calculating our project velocity with some spreadsheet calculations; we had no automatic way to select which stories we would be able to work on in the next week. The automatic velocity tracking completely changes that.

Pivotal Tracker itself is not Free Software. While we use it, we trade away some freedom for that efficiency. And because it runs on Pivotal Labs’ servers, I don’t have control over it. I am subject to the whims of Pivotal Labs, who owes me no particular guarantee that it will remain functioning or keep our data intact.

To help mitigate the damage from data loss, I’ve already configured a job on our server to back up our data nightly. That way, if Pivotal Tracker collapses, we only lose one day’s work. You are free to use  “Pivotal Backup” yourself – our backup tool is Free Software. This is possible because Pivotal Tracker does not lock us in. It offers complete export and import of its data. Furthermore, the terms of service seem to allow me to use this “bot” to automate the process.

Right now, if we refuse to use Pivotal Tracker for freedom reasons, we would have to write our own tool that did the same thing. By using PT today, we delay writing that tool — perhaps we’ll never have to write it at all. And if Pivotal Tracker goes away abruptly, we can immediately go back to using a shared spreadsheet of some kind — PT exports our data in CSV, which all spreadsheets import. For tracking tasks,  Pivotal Tracker replaces our use of Google Docs’ spreadsheet sharing. We were already relying on a non-free package; PT is just more helpful.

I personally remain uneasy about our use of non-free software at OpenHatch. Our purpose is to advance open source and Free Software. And personally,  I agree with the autonomo.us movement emphasizing Free network services.

But this time, I can see that Pivotal Tracker makes us more efficient. Our goal is to make open source and Free Software better; using Pivotal Tracker helps us do that faster.

For now, I’ll accept the pragmatic approach that Mike Linksvayer taught me when I worked at Creative Commons. When CC transitioned from hosted Free Software email to mail hosted by Google, I voiced concern. Mike pointed out that, when CC wanted to, it could always switch; CC controls its domain name, so if Google ever became abusive, CC could quickly switch to a different mail hosting service. Besides, Google supported data export; CC could pull the email out of Google’s servers and drop the messages somewhere else. CC was the final arbiter of who controlled creativecommons.org email, even if for today it gave up that control to Google.

For us, with our backups, we maintain that relative degree of control.

Going forward, we’d love to use a Free alternative. Can you help us find one with this degree of simplicity and automation? And maybe we can all convince Pivotal Labs to Free Pivotal Tracker.

7 comments

  1. Simone says:

    Yes Pivotal Tracker is really well done and we are using it for all our projects now.
    One thing it was missing was a “Pin Board” view of the current iteration but then Google Chrome saved my day.

    I developed an extension to do just that

    https://chrome.google.com/extensions/detail/iegbkljacgfochoondhgibofiijnedjd

    check it out!

    vizio

  2. Lars Nordin says:

    Are you aware that there is a Python lib for Pivotal’s REST interface?
    See Pytracker at http://code.google.com/p/pytracker/.

  3. Asheesh says:

    Lars, thanks for that note! As far as I can read at https://www.pivotaltracker.com/help/api the backup features still aren’t part of the API. So we still have to screen-scrape to get the nightly snapshot. (And that’s what “Pivotal Backup” does.)

  4. Niall says:

    Just a thought: the backup time and amount of data is probably tiny. Would it make sense to change the backup frequency to e.g. hourly?

  5. Mizery De Aria says:

    Are there any updates to this blog entry now that Pivotal Tracker is set to charge for service?

    • Asheesh says:

      The only real thing I can add, right now, is that we (OpenHatch) has stopped using Pivotal Tracker. Roundup, our self-hosted bug tracker, provided email integration features that we really like, and we could skin it to match our community.

      The planning features would be nice, but as we’ve transitioned into a volunteer-based project, I’m wary of telling people that our velocity is “too low”.

Write a comment