2 ms·
> To me, with a timeline, livestreaming and archiving is the ultimate in source control. What if you could trace a line of code to a check in to a time stamp to
by minimaxir 12y ago
> To me, with a timeline, livestreaming and archiving is the ultimate in source control. What if you could trace a line of code to a check in to a time stamp to the video of that line of code being created? What if you could stitch that all together to a work item or ticket, and get video of the design and thinking process, to understanding why the code is a certain way? That's exciting.
Isn't that the exact use case for code/commit comments? :P
- tree_of_item 12y agoCommits don't even come close to capturing the amount of data a screen recording does.
- minimaxir 12y agoTherein lies the problem. Screen recordings aren't a concise representation of the programmer's thought processes, and it would take any viewer more time and effort to watch 15 minutes of video for a problem context and resolution than a concise 1-2 line comment summary of the code delta.
- ColinDabritz 12y agoCertainly! The compression is handy when you have it. This is why check-in comments like "Fixed bug #1234" are really frustrating. The link to the bug is nice, but out of a list of 10 similar comments, I need to find the one that say, turned on the security. I need the compression. That said, sometimes my goal isn't "find it quick", it's "understand it fully". Perhaps we'll figure out ways to understand the video in fast forward, make visualization of code change timelines, that sort of thing. But I think it's interesting to start with all the raw data and figure it out from there.
- ColinDabritz 12y agoQuite, but this would be moment by moment, not "huge chunk after an hours work", a proper stream of the flow of time, and very high fidelity. Isn't source control just a fancy way of saving plain text files? In a certain sense sure, but it's also a lot better. I think it would be interesting to explore. I think the big difference for me is treating a timeline as a first class programming artifact, giving you locality on the code, and being able to explore. What lead up to this code? How long did it take to create? What dead ends were explored on the way? etc.