ChoosingAppropriateTags

TaskEstimation uses tags to allow filtering of tasks and task breakdown. To facilitate dataset filtering and produce estimates faster, it's crutial that the tasks in your data pool are tagged properly. It will allow for faster filtering and testing of different alternatives. This page describes the type of tags that can be relevant to your estimations. However, this list is not complete and open to experimentation.

Tagging Tasks

Tags placed on the tasks will be used when filtering the historical dataset when an estimate is required. The best datasets are those in which the data contained is the most similar to the task you are attempting to estimate. The most relevant that can be used is the one from the same project. If the project has been running for a few months and you attempt to perform an estimate on an new task, the historical data from the project alone can be used. If this is not the case, you may fall back to projects using similar technologies or a similar application domain. The further away, the less precision should be expected.

At the moment you are tagging the project, you likely do not know in which estimates it will be used, so the more information, the better. You definitely want to avoid having to go through your task history to add more tags. Tags must characterize not only the project itself, but any other attribute that can affect development time, like the quality requirements of the project.

Here is an initial list you can use to build your tag list:

• Project code
• Client or customer
• Technologies used
• Project environment (web, intranet, extranet, embeded, ...)
• Target userbase order of magnitude (ub1, ub10, ub100, ub1k, ub10k, ...)
• Team size

Once you have sufficient amounts of data, experiment with your datasets to figure out which factors provide the best correlations between your task size and time to accomplish.

Tagging Task Breakdown

Wether you choose to use the subtask breakdown or the spacial breakdown, it's possible to tag each of the subitems. Unlike tagging tasks, these tags will mostly be used in a near future: when performing the estimate and after completing the task to review the estimation work. The quantity is not as important. They will mostly be useful to separate the different aspects of your task and use different datasets to estimate each of them. For example, a modification to the database schema may require a one time only script to convert data, update of the code libraries accessing the schema and creation of a new user interface to access the new information.

These tags will help in identifying the aspects of the tasks that hold the most weight.

Simply use the tags you intend on filtering with.

---