Hacker News Reader

Comment by ajarmst | Hacker News Reader

No command history, no tab completion. One assumes that "shell" is aspirational rather than descriptive. It's an interesting idea, but if you present what appears to be a shell interface to users expecting a shell interface, they're going to wander away pretty quickly after discovering it's merely facade. I wish you luck, but this isn't really ready for visitors yet.

View on HN

It is totally ready for visitors. I visited, I enjoyed it. Doesn’t seem particularly wise to advise people something isn’t ready for people to visit, when those visitors are the only ones who can actually give useful feedback.

One hopes that the author didn't need me, a stranger, to tell them that this site is missing most of the fundamental features of a shell. Feedback is valuable, but strangers aren't going to visit something that clearly doesn't work a second time to see if they're making progress. There is a reason that early testing and feedback on prototypes is limited to in-house or pre-existing customer review. The author is entitled, and even encouraged, to ignore me--but my feedback is that they should have waited until they were closer to their goal and general feedback would be more valuable than hundreds of copies of "it doesn't work".

> but my feedback is that they should have waited until they were closer to their goal and general feedback would be more valuable than hundreds of copies of "it doesn't work".

This is the exact opposite of useful feedback. I hope you realise project that is literally intended for fun...

What definition of "shell" are you basing this on? Some (current) Unix systems I use have default shells that have neither command history nor tab completion.

Easy access to command history and tab completion are certainly niceties that came along at some point, but are in no way part of the definition of a shell.

ok, I'll bite. What current Unix systems ship with a default shell that supports neither command-completion nor command-history?

The one I have to use frequently is the restricted ksh environment in IBM's IVM/VIOS server, which is AIX.

I believe it does log a short set of command history to a file, but it'd not available with, say, the up arrow.

In general, though, I don't think any form of history is required by the POSIX shell standard, and tab-completion definitely isn't.

ksh93 supports both completion and history. I lost interest in looking before uncovering information about versions of ksh more than thirty years old, but neither feature is mentioned in descriptions of differences between ksh93 and predecessors, so I suspect the feature has been there for longer. Posix compliance is neither necessary nor sufficient for a reasonable colloquial definition of "shell."

hi thanks for the input, yes it is very early and I just posted this mainly for validation; I wanted to know if only I thought this was cool or if others also saw it as something fun. I did not anticipate so many people using this. I plan to implement most of the expected shell features soon.

Cool. It's probably not an original idea, but it's good one. I might even steal it from you at some point. Good luck with the project.

This is a personal website more for showing off than for being useful.

That the quality of someone's work doesn't matter because they don't care about it or expect it to be useful isn't a particularly compelling argument.

I think you’re being overly critical of someone’s hobby project. This isn’t an enterprise app. It’s a fun little thing they made and wanted to share with us. If this is how the community reacts to what is really barely a hobby project it doesn’t surprise me at all that devs are burnt out from supporting OSS projects that probably have 1000s of ppl like you that complain and expect the world for free.

I thought it was pretty cool, and glad they shared it.

Reply on news.ycombinator.com