This is a mobile version, full one is here.
13 May 2015
Two Instruments of a Software Architect
A software architect is a key person in any software project, no matter how big or small it is. An architect is personally responsible for the technical outcome of the entire team. A good architect knows what needs to be done and how it's going to be done, both architecturally and design-wise. In order to enforce this idea in practice, an architect uses two instruments: bugs and reviews.
At Zerocracy, we discourage any communication between developers unless they are formally attached to the tickets or tasks we're working on. Read more details about this approach in this post.
The same principle applies to an architect. We don't use meetings, stand-ups, Skype calls, IRC channels, or any other tools where information flies in the air and stays in our heads. Instead, we put everything in writing and talk only when we're being explicitly asked to and paid to—in tickets.
With this in mind, a reasonable question may be asked: How can a software architect enforce his or her technical vision for the team if he can't communicate with the team? Here is our answer: the architect must use bugs.
A bug is a ticket that has a reporter, a problem, and a resolver, just like this post explains. Say an architect reviews an existing technical solution and finds something that contradicts his vision. When such a contradiction is found, it is a good candidate for a bug. Sometimes there is just not enough information in the code yet, and this is also a good candidate for a bug.
Thus, bugs reported by an architect serve as communication channels between him and the team. An architect doesn't explain what needs to be done but asks the team to fix the product in a way he thinks is right. If the ticket resolver, a member of the team, disagrees with that approach, a discussion starts right in the ticket.
Sometimes an architect has doubts and needs to discuss a few possible solutions with the team or simply collect opinions. Again, we use bugs for that. But these bugs don't report problems in the source code; instead, they complain about incomplete documentation. For example, say an architect doesn't know which database to use, MongoDB or Cassandra, and needs more information about their pros and cons. A bug will sound like "our design documentation doesn't have a detailed comparison of existing NoSQL databases; please fix it." Whoever is assigned to this ticket will perform the comparison and update the documentation.
Bugs are a proactive tool for an architect. Through reporting bugs, an architect influences the project and "dictates his will."
In our projects, every ticket is implemented in its own branch. When implementation is done, all tickets pass mandatory code peer review. In other words, developers review each others' code. An architect is not involved in this process.
But when peer review is done, each ticket goes to an architect and he
has to give a final "OK" before the code goes to the
master branch through
Rultor, our merge bot.
This is an architect's opportunity for control. This is where he can prevent his vision from being destroyed. When the code created by a developer violates project design principles or any part of the entire technical idea, the architect says "No" and the branch is rejected.
Reviews are a reactive instrument for an architect. Through strict and non-compromising code reviews, an architect enforces his design and architectural principles.
PS. Here is how an architect is supposed to report to the project manager: Three Things I Expect From a Software Architect