Response to Solderpunk's suggestion to use Git

This post is in response to solderpunk's gemlog post that can be found at the below link

Low Budget p2p Contint Distribution with Git

As a note: I do not use git. I have had no real reason to USE git before beyond 'go clone this git, follow these instructions, and your whatsit should be up and going.' Which was helpful in getting proper wifi up on my machine.

I'm on my second read through since.... just at a concept level this gets a 'but why?' out of me. Ya, jumping to a whole new protocol is neat and exciting and has some uses. Re-purposing git has me scratching my head. In a way I get it at a conceptual level and like it. Have the client do a pull request for your material so that if something ever happens to yoru server their copy still exists. Which part of me finds agreeable.

On the other hand I dislike how Solderpunk is talking that the pull requests seem to be automated and continual whenever you visit a given site. What if I stumble on a site i don't want? I shouldn't have to go dig through settings to kick it out of the pull request chain. Nitpicky I suppose and easily fixed by making it a proactive choice to add a given site to your pull request list (Essentially your favorites list really.) The other is why not automate the process? I'm fairly sure that is something git already does and allows since I can literally go into command line and 'git clone: [address]' So having a timer function to make pull requests at x time and or when connect next' should be a no-brainer for clients to handle.

I don't like this as a solution though since it feels like one of those things that works just well enough to fix the problem, but doesn't solve the other problems Solderpunk himself admits this has yet keeps talking like it's 'temporary.' Software, much like carpentry, is a case where the temporary fix often becomes permanent.

So while conceptually having a git-like setup for gemini servers to use so clients with the git feature (or just users performing pull requests) is neat. The fact that the guy that is pitching this as a fix himself admits this doens't solve everything leaves me not liking it, since any fix down the road will have to account for this blessed by the creator concept is now out there and ... I mean there are worse things than having background pull requests that are handled by the client? I'm just afraid, as a user, that this could introduce something that would hamstring a more proper solution rather than this '80/20' that Solderpunk is speaking of.

-_-_-_-_-

Gemlog

Go Home