Writing, technically | Hacker News Reader

Writing, technically

signalsandthreads.com

135 pointsfezzez16 comments

The intro touches on something I learned to be true very early in my technical writing career, but which isn't discussed much and may not be widely known:

> helping to push forward all sorts of efforts around knowledge-sharing at Jane Street

As a TW I often find myself advocating for more scalable/sustainable/permanent ways to store institutional knowledge. The classic example is email. Sometimes there is extremely useful knowledge buried in an email thread. I often work with engineers (or whoever) to turn that knowledge into publicly searchable documentation.

P.S. tangentially related to the title of this post, my brother came up with a fun self-deprecating joke (which I later used on as a tagline [1])

> Technically, I'm a writer

[1] https://developer.chrome.com/blog/welcome/

Every company I work for struggles with knowledge capture/sharing, and it’s always a priority, but never really prioritized. Tough nut to crack.

It is especially true because it is much faster to write an email than to create a piece of documentation. Even more because of the state of corporate knowledge sharing tools.

But yeah at the end of the day, there is no easy fix, it has to be in the core list of responsibilities of each job to work at the company scale

Same. Unless you make it your personal responsibility, it won't get done. Documentation is hard, and not everyone's definition of fun. It's especially hard to explain things you are intimately familiar with to your colleagues.

I've done that for long enough to be very good at it. Now I earn my living from documenting stuff the local government doesn't document property.

So true.

People don't realize and/or don't care much about all the information that gets lost in e-mail.

I've been struggling for years to push people to document stuff. Often I'm the only one who writes page after page, and send links to docs, until there are hundreds. At that point, others usually start to chime in a bit.

At my day job, they delete email after 180 days. I often wonder how much time has been lost recreating information that was deleted that way.

This happened at my last company, but with Slack. It was a real pain at times.

And how much money might have been saved not having some of that information available later... because that is totally a good reason to burn history.

I listened to some of this today and it's really great. Lots of similar learnings as has been had from the g3doc team at Google.

More information available here: https://www.youtube.com/watch?v=EnB8GtPuauw

I think a lot of companies would benefit immensely if they hired a few tech writers. Management always talks about collaboration and knowledge sharing but they refuse to put money behind this desire. Instead it’s up to the workers to create documentation while also having to finish project deadlines. And most developers don’t really enjoy writing or are good at it.

TWs help for sure but the ones with the ones with the most impact don't try to do it all themselves, and instead make it easier / less painful for everyone in the company (or at least engineers to start) to create documentation. Or they at least raise hell and make it clear that the more everyone takes ownership of docs the more likely it is that the company overall will succeed. To put it another way, a common problem that I see is that company's recognize their docs are terrible and then hire a couple TWs and the engineers say "phew! Now I really don't need to care about documentation at all!" I think part of Stripe's initial success was that everyone supposedly did customer support and that helped them see that good docs was a great way to reduce support burden (rather than providing ad hoc answers, if you have a doc that explains the solution well, you can just link to that instead). That's just conjecture, BTW. I haven't confirmed that idea with Stripe people. Also it's just common sense that if your main product is an API then your documentation is mission critical to overall success. Another scalable approach is for TWs to provide training that helps non-TWs get less intimidated by writing docs.

"make it easier / less painful for everyone in the company (or at least engineers to start) to create documentation. "

Exactly. I am not averse to writing most of the content but at least in my company the documentation processes are very intricate and I really don't want to learn all the details and rules for submitting docs. I have enough to do with wrestling with AWS and feel my time is better spent on that.

Same for booking flights. We used to have somebody who would book flights. That role got eliminated to save costs and now the engineers are burning countless hours figuring out how Concur works and what the latest travel rules are.

It's nice so have support staff that are experts in these systems. And I think it's actually cheaper. there is a reason why the big guys have assistants.

We had some success making it easy for engineers.

User documentation was inline with the code (think perldoc/pydoc). Developer documentation (as comments) was also co-located with the relevant code.

This automatically gave us man pages at a shell prompt. We could include those man page outputs in report .zip files that got emailed to users, and push them to our website.

There's an art to keeping this organized. Always having a good trail of breadcrumbs was important, like a comment next to a variable assignment telling the developer to update the user docs if that constant is changed. Also avoiding duplications, which leads to drift/inaccuracy. We considered inaccurate documentation to be a bug. Full time employees took to this. Contractors often needed reminding.

As the project grew there was managerial pressures to get involved in other stuff like power points and sharepoint.

I have seen several of such initiatives. They all started out well but then slowly died because management didn’t allocate long term attention, time and budget. If this is not part of the fundamental company process it will die slowly.

For those that didn't listen to season 1, I recommend it. Really interesting show.

This is so helpful

Reply on news.ycombinator.com