Hacker News Reader

Comment by zdgman | Hacker News Reader

I would love to hear why the github team waited so long to bring this feature into existence? I am sure there is a good reason (or maybe not) but it would be awesome if a member of the team would stop by and post the details.

View on HN

We don't have roadmaps or a prioritized backlog, so we weren't exactly "waiting" to implement this. Someone got fed up with the image workflow, figured out a solution, grabbed some other people to help, and shipped it.

can you speak more to how this is built - even if its just at a high level?

Someone came up with a design. Someone else implemented the backend. We iterated on it for a bit and shipped it internally. Then we figured out the CDN situation, setup logging and metrics, and shipped it genpop today.

Can you tell me a bit about the CDN situation, especially what specific issue did you run into?

No issue, we just wanted to set something up as opposed to running directly from S3.

Ah nice!, easier to drive down the costs that way. Thanks.

Now I know github doesn't have a typical roadmap or backlog but I guess what I was looking for was how someone decides what feature they are "fed up" with not having.

We use GitHub to build GitHub, so typically a feature comes from our own usage pain points.

I imagine the conversation internally went something like, "Man, attaching an image to an issue is really annoying. Let's make it not suck." An issue and pull request followed, internally shipped it to test it, and today rolled it out to you all.

That's literally how 99.99% of features happen on GitHub. :)

Thanks Jeremy. That's the sort of answer I was looking for. Good luck with Railcar!

Pain points rule everything around me.

I was working on a project where, when viewing an issue, you could use the file picker to choose a file that was then committed on the project's 'gh-attachments' branch, and then linked via markdown to the raw URL.

This is good enough for me :)

Reply on news.ycombinator.com