Fog Creek Software
Discussion Board

Ginsu task management knife?

I work in a mid-sized (~80-person) IT dept., facing the same organizational probs I've seen time & time again. We have both project and ad-hoc ("support") work. We work for 'clients' -- other company divisions -- and sometimes we're our own client. We have separate depts. for project mgmt, project-related development, data mgmt, ad-hoc-related development, infrastructure (servers, network, desktop, etc.).

We need to track tasks for a variety of reasons:
We break down projects into tasks. Some of the tasks go to outside consultants. Some of the tasks got ot project-oriented developers. Some go to support techs.

The tasks occur at various stages of the software lifecycle: a project is broken into tasks so we can make a plan. Support work is generally parceled out as discrete tasks. During UAT, reported bugs become tasks that need to be sent back to the project developers; after rollout, bugs become support tasks for the support developers.

We need to at least track the costs of these tasks, and in many cases we need to allocate or bill back the costs, so we need to categorize by various criteria.

I'm looking for a good set of tools to organize and track all these tasks. Currently, we have several, and they don't talk to one another very well:
1) a meta-project-management app, which tracks all the projects.
2) MS Project, which we use on an individual project. Tasks for project developers come straight from Project.
3) Magic, a Help Desk app that we use for support tasking. Project tasks that go to support techs get recreated here -- recreated being a sorepoint, cuz there's no linkage between the apps.
4) bug-tracking is not organized -- each project does it differently, with yellow pads to whiteboards to Excel lists to Access DBs to Web front-ends (no FogBUGZ yet ;-).

I'm looking for a process and supporting toolset that does a better job at integrating these various assignments. Should we be tasking everything out of MS Project, and make the support folks check there instead of Magic for what to work on next? Conversely, should we convert every project task into a Magic ticket, and make project folks live by the Help Desk queue? Are there products that do a better job at integrating these things? I *know* I'm not alone here -- I've faced the same problems when in IT and in consulting.

BTW, I'm NOT asking about whether we have an ideal IT organization -- it's a bit messed up, but not that different from others I've seen. I haven't had the pleasure of working in a really well-organized company yet! So slagging about how f*'d up our shop is misses the point -- we know that, and even if we can fix those problems, I still want good tools support.


Tuesday, July 29, 2003

Have you asked these questions on one of the project management fourms that exist on the World Wide Web? How about on Usenet?  If so, did you get any feedback?

One Programmer's Opinion
Tuesday, July 29, 2003

Why not roll your own solution, if you can't get what you need out of what's already available?  Sounds like you spend a lot of time fiddling with what you have, why not devote that time to putting something together that will fit your needs precisely?

Tuesday, July 29, 2003

I think you can pretty much handle 90% of it with Fogbugz - do the trial thingy and have an open mind about creating categories and projects and things.


Tuesday, July 29, 2003

Maybe I am missing something, but the tracking of all of this should be accomplished in MS Project.  The exception is detail in your magic and bug tracking. 

Consolidate on a bug/issue/support tracking system (we use ClearQuest, but there are dozens of good ones).  Then tie the task back to Project. 

The issues I see are when people decide not to tie support back to a project.  The "it's in production so everything is now a support issue" hide the true cost.  Especially on projects were "we made the date" but 30% of the functionality is missing.

Tuesday, July 29, 2003

Seriously consider StarTeam. It can handle the issues you describe.

I'm a QA Manager and have used StarTeam for several years, so I speak from the end-user/customer perspective. I do not now nor have I ever worked for StarTeam or Borland, and have no vested interest in recommending StarTeam.

If you search through previous threads here, there have been some active discussions about it, many going into pretty good detail about its features, so you should be able to get plenty of info about it. Drop me an email if you care to hear more about our experiences with it.

If your budget allows, go for the gold and get the top-end enterprise version, because with that you get the ability to reconfigure interfaces (they have a facility whereby you can virtually replace the existing UI with a swing-based version of your own). One peeve of mine has been that I want the interface for entering CR's ("change requests", i.e. 'bugs') to be different. The enterprise version would permit that.

I'm not blind to it's quirks, but I recommend it without reservation. It's a good tool for large groups as well as small. I have personal knowledge of at least one extremely large and well-known corporation that is putting major projects 'into' starteam; including some whose teams are located on two continents (tens of thousands of employees world-wide; though most will not be 'on' or using starteam itself). StarTeam would easily handle the size organization you mention.

I've been around a pretty fair variety of development organizations, both commercial "shrinkwrap" as well as consulting/project-based. I understand the needs and issues that cause organizations to buy tools for project mgmt, bug-tracking, help-desk/trouble-ticket, CM, deployment, and collaboration. Having lived with both discrete tools and StarTeam's integrated capabilities, I would not want to go back to using discrete tools.

Obviously, I'm sold on StarTeam, but it's not without a good basis of experience, nor as I mentioned before, a blind eye to areas where it could be better. It is a very powerful tool with a broad scope of capabilities; as such there's a fairly steep learning curve to successfully employing it in your organization. This is especially true if you've got people who don't understand anything about CM, for instance (a more common problem than you might realize, BTW). Of course people who don't understand CM and SDLC will have difficulties regardless of the tool used, because they don't yet understand the actual problems these tools exist to solve. So, you'll have to assess your people -- you may need to do some 'basic' education about the software development life cycle anyway, which wouldn't hurt even if you chose no specific tool at all.

On that note, a word of caution--ignore it if it doesn't apply to your situation--No tool, by itself, can fix an organization with deep-seated systemic problems. I've seen this sort of thing too many times -- somebody gets all hot and steamy about a new tool or methodology, convinces the right person, and then tosses it out to the team, but to little or no positive effect (maybe you've heard the phrase 'rearranging deck chairs on the Titanic'?). The right tool or tools can certainly help an organization do particular things much more easily and accurately than they could without them, but take care that you're actually solving the right problems first.

StarTeam provides powerful capabilities to organize and manage your development efforts. It's definitely worth a good look. If you don't go with StarTeam, I still strongly recommend an integrated system -- way better than trying to trace, establish, and maintain the logical relationships among tasks, changes, requirements, and versions manually from among a "Frankenstein" collection of discrete tools.

Good luck,

starteam user
Tuesday, July 29, 2003

*  Recent Topics

*  Fog Creek Home