Anmelden / Registrieren Konto

One of the most common issues seen with experienced professional software developers who start to work on community software is a reluctance to engage with public communication channels, like mailing lists. Understanding the reasons why, and helping your developers engage with the community, is a key to creating a successful and fruitful relationship with the community you are working with.

In my experience, some common reasons for this reluctance include a lack of confidence in written English skills, a perceived lack of technical skills, nervousness related to public peer review, and perceiving community interaction as "communication" or "marketing," which they believe is not part of their job.

Lack of confidence

Traditionally, university engineering courses do not put a strong focus on clear written communications. This can lead to professional software developers who have not developed their writing skills. In open source projects, written communication is critical to working asynchronously with a community team:  blogging to promote work, documenting feature proposals, or debating the pros and cons of roadmap items on a mailing list. 

In addition, while most engineers appreciate good logic, rhetoric and debate are not typically part of an engineering syllabus. The end result is that when asked to write about their work, or to ask a question on a public forum, they hesitate.

For non-English speakers, lack of confidence in English proficiency can also be an issue that gets in the way of mailing list participation.

I remember one engineer I worked with, who was very hesitant to write email about topics we discussed, even when we agreed that they needed to be discussed with the community. When he did, he appeared to ignore responses or questions from community members. 

When this was discussed with him, it was clear that he did not feel comfortable engaging in a discussion in writing on a publicly archived mailing list. He was nervous that any mis-steps or breaches of etiquette would be dealt with harshly by the community.

The best way to cure this problem is with baby steps.

A great first step is to get engineers answering questions. One way to help is to publicly refer a question to an engineer on the mailing list, while affirming their expertise in the area. "Tammy knows the RLE algorithms inside out – Tammy, would you mind answering Franz’s question, please?"

This is something of a Jedi mind trick, and serves three purposes. First, an individual can easily ignore a question if it is addressed to a group, but it is a rare person who doesn’t answer when you ask them a question directly–it’s human nature. 

Second, affirming your engineer’s skills reinforces that they are indeed seen as the expert in the area in question, giving them more confidence to answer. Third, by identifying an individual engineer with a specific skill set, you are giving people outside your company a glimpse inside the walls. You are making your team human, a collection of individuals with different strengths and weaknesses, rather than an amorphous group – "the Acme Co developers."

Another worthwhile thing to do is to get developers into the habit of writing regularly. It is not enough to stand over people and ask them all to have a blog. If writing is intimidating, then doing it more often will make it less so. There are many ways to do this–rewarding blog posts, requiring regular status reports, longer commit messages or comments when closing tickets, or scheduling time for creative writing workshops. The goal is not to turn developers into novelists. The goal is to get your team in the habit of writing.

Finally, you should train your engineers in basic netiquette and writing good emails. There are lots of good resources out there, including Producing Open Source Software by Karl Fogel, The Open Source Way, the Linux Kernel Contribution Guidelines, and Tips for asking technical questions that result in fast, useful solutions

One suggestion I have heard is that developers should treat writing email in a similar way to patches. When you generate a patch, typically the last thing you do before you send it is you check over it, to make sure nothing silly is included. The same habit applied to email would identify any places where phrasing is awkward and ambiguous, resulting in better email.

The peer review gauntlet

While writing can be intimidating for many software engineers, subjecting their work to peer review to a group of people they do not know very well can be nerve-wracking.

In my experience, systematic peer review is not the norm in the software industry. Some managers see peer review as overhead. After all, the developer was hired because they were competent to do the job, and nobody likes to be second-guessed by "the community." Once engineers reach a certain level of experience, peer review seems to be more an exception than the rule for professional software developers in our industry.

In community projects, peer review is expected. In fact, it is a best practice, one of the things that separates successful community projects from the crowd. Community developers expect to hear about features before they are developed, and have an opportunity to suggest better ways the feature can be implemented. They expect new contributors to submit patches that they can review–it is the way a new contributor builds trust before gaining committer or maintainer status.

The best way to get people used to peer review inside professional software teams is to have a company policy against the "day one commit bit"–the practice of getting commit access to an open source project repository on the day you start in the company. For corporate-sponsored projects, new developers should go through the same review process for their work that contributors outside your company have to go through.

For corporate contributions to community projects, that means discouraging internal branches and a "gatekeeper" project structure, where one or two developers commit the work of others in the team. 

Developers should submit their work upstream at the same time it is being submitted internally. For those changes which only apply to your internal branches, peer review should still occur in a private repository. With the recent advent of git-based repositories as a norm for software development, this is less of a challenge for companies now than it once was, and it is necessary to transition your team to successful community contributors.

Having new developers go through this review period is important for a number of reasons–the most important is that you are demonstrating that employees have the same burden to prove themselves in the project as non-employee contributors, and it provides new employees with a period where they familiarise themselves with project coding and communication conventions and norms, and when they also introduce themselves to the community at large. This is fundamental to the mentorship of new developers by more experienced community developers.

Communication and marketing are not my job

In the mind of some developers, posting project plans to the mailing list constitutes an announcement and needs substantial preparation. These developers frame communicating with community users on a mailing list as being equivalent to "support" or "communication," and not as a core part of the engineering function.

If a developer frames "sending an email to a mailing list" as "making an announcement," it fits in the "marketing" box in their head. A developer once told me she didn’t have time to send email to the mailing list, because she had real work to do. "Dealing with the community" was my job as community manager, as far as she was concerned. In her mind, community messaging was a type of support, and was not part of her job.

The authors of The Cluetrain Manifesto claimed that in the modern connected world, there is no such thing as a marketing department, that every interaction between an employee of your company and someone outside your company is an opportunity to win or lose reputation.

In free software development teams, this is even more true. There is no marketing department for project communications. To be productive, you need to talk to your peers, so the institutional barrier to external communications must disappear for people dealing in community projects.

Even in organizations that are clear about their expectations for open source communication, engineers may impose limits on themselves. They might spend hours polishing proposals before sending them on. This is counter-productive in community projects, since the more polished an initial proposal is, the more emotional investment has been made in it. Polished proposals are harder to review and change, too, since they look done already. It is better to release a rough early draft, giving the author an opportunity to integrate early feedback.

The best way to prove to your team the benefits of "release early, release often" is by example. As a team lead or manager, you can lead by publishing team plans publicly and early, and iterating often. When you do, point out the benefits which result to your team. Breaking down this barrier will take time, but by making it clear that perfection is not expected, and by rewarding early release of information and encouraging feedback, your team will soon learn that participants outside your company are peers, not an audience.

Another useful technique is to ask your engineers to break tasks into smaller parts, so that even if they do hold off until the first part is up to a high standard, information is still getting out there more quickly, allowing feedback to inform later stages of the process.

Working against the goal of early communication is the common desire for the big reveal. Companies often want to align product releases and announcements with major trade shows. This can lead companies to ask their engineers to work internally on significant features for fear that the big surprise will be ruined otherwise. The alternative seems to be to announce a project when you start, rather than when you have something to show–but this can result in a long wait before products get to market, and impatience and bad press from the mainstream press.

It is possible to separate engineers discussing design decisions and implementation details of significant features in a mailing list, and using a press release and marketing campaign to promote an announcement. When the final announcement comes out, the community will not be completely surprised, and you will not find yourself having to defend yourself for working in secret for months, and proposing a big code drop which is difficult to review.

Building relationships

The key lesson here is that you want your developers to feel a connection to people working outside the company. That requires people outside your company to feel a connection with them, too. By drawing the curtain back on your team, its members and their skills and priorities, you are creating the circumstances for the people working on your project to come to appreciate each other as peers, and to feel comfortable discussing features and patches on their merits. When you get to that point, you have won the battle. Developers in your team will see other developers working on the project as peers, colleagues, and even friends.


Highlights

Das könnte für Sie von Interesse sein