Fork me on GitHub
#off-topic
<
2019-06-18
>
Stefan06:06:48

@jayzawrotny I like the article, I don’t have any particular feedback. Your question reminds me of a tool that I’d like to share in this context: Hemingway Editor. It’s a tool that I have used in the past to get some feedback on my language. It grades the readability of your text overall and even on a sentence level. http://www.hemingwayapp.com (I’m not affiliated to it in any way, just think it’s a nice tool 🙂. )

jaide15:06:03

Thanks! That’s a great way to improve my writing quality.

Olical16:06:53

For versioned software developers or consumers (presumably OSS): Could you see yourself using a tool that lets you publish structured changelogs (which could be converted to plain text and placed in your repo / GitHub releases) etc? You'd get to see stats on what versions people are using and how they're finding upgrading. From the consumer perspective you can subscribe to those changes and be notified by email with one click skip / snooze. All changes from your current version to the desired version would be bunched up so you can read them all together, breaking changes would be highlighted so you can work through them carefully. Basically leading to lower stress monitoring of high risk dependencies, knowing that you're not going to miss something important buried in a .md file. (feedback on this concept appreciated as a thread!)

kszabo16:06:47

This sounds interesting to me, but as Rich outlined in his speculation talks I would rather have dependencies which don’t break stuff at all 🙂 Also the problem with human-defined changelogs is that different people think differently about what a breaking change is, leading to the usual semver problem that a patch release breaks consumers. See: https://imgs.xkcd.com/comics/workflow_2x.png

Olical16:06:59

Of course, that's the idea. Right now I'm causing breaking changes a lot because I've got some users of #conjure (Neovim prepl tooling) but it's still early development, pre v1. So I'm breaking things and trying to help people keep up as I stabalise it. Even with something that never broke, it would still give you a nice UI for batching up lists of things that were added and news about what things you should probably stop using, even if they're not going away.

Olical16:06:56

Just trying to gauge if the fundamental idea is a non starter right now really. Like if it sounds dumb to others. I think I'd enjoy building it!

Olical16:06:25

Would be cool to see stats on how break-y a tool is too. Would be a fun badge to add to your readme. The stats that I could give people out of this tool could be really handy.

kszabo16:06:17

yeah, that pitch sounds better. The problem is going to be adoption if you go the human-defined changelog way. For instance I’m really interested about some perf toggles that Jetty might introduce, but if they don’t use your product then the added benefit that you would provide is lost. I would rather scrape data of existing outlets/blogposts/changelog.mds first and hand-curate stuff that people want. Ie. people give you deps.edn/build.boot/project.clj dependency description and do a top-n most-requested dep prototype

Olical16:06:12

Yep, totally get that. It would be something maintainers would have to push which would give you a (hopefully) nice UI to link to from your readme but it'd also maintain a text file in your repo. It wouldn't be something consumers can pull initially, maintainers have to push willingly. Scraping and parsing is something that it could maybe do, but wouldn't be an MVP feature if you get what I mean.

Olical16:06:37

MVP is "I have a place to publish releases that helps people and sends notifications". Then I build upon that concept over time.

kszabo16:06:33

I say go for it 🙂

kszabo16:06:03

I think a model like http://cljdoc.org would work nicely

Olical16:06:28

Some things auto but enhanced by user input?

kszabo16:06:42

yeah, if they already use github releases/changelog.md just try to use that

kszabo16:06:51

but make it editable through github auth for them

Olical16:06:14

Yep, I was musing about trying to make it work without input but much better when you start to create your change log through it since it has more context.

Olical16:06:18

Food for thought!

Olical16:06:51

Just looking to see if people reply with "why would you do this nobody needs it" right now really 😅

CyberSapiens9716:06:24

do we have a job/role name, for the guy who Designs/Build the infrastructure of a particular system? by infrastructure i mean all of it, the software architecture, how the machines are architected in the system as a whole and so on...