Hacker News Reader

Comment by LarryClapp | Hacker News Reader

Thanks!

Yeah, I've looked at warp.dev. Funny story: I heard about them from Daniel Martí, aka mvdan, the author of the shell library I'm using.

It looks really neat, though I confess I haven't explored it much; dogfooding, don't'cha know. :) I felt a little threatened by them but also weirdly encouraged, since they got $50M in funding. It made me think that hey, somebody thinks a new experimental shell/terminal is a viable commercial product; why not mine? Also, Warp is pretty expensive by comparison (for teams of 6 or more, anyway; granted it's free for 5 or less).

Yeah, getting people to switch will be difficult, especially since most terminals are pretty rock-solid as far as their actual ncurses emulation, and mine still has some rough edges. I'm hoping that I can at least get a few happy users and go from there.

Thanks again for the kind words!

View on HN

somebody thinks a new experimental shell/terminal is a viable commercial product; why not mine?

I'm glad people are working on these although as you say yourself, commercial viability is tricky. I like the non-SaaSyness of this but then I'm still left with having to:

1. change my shell

2. change my terminal

3. but not on Windows

4. rely on a single person for bugfixes, security issues, features

5. pay $40/yr for the privilege

None of these things are individually an insurmountable hurdle but they add up to a pile of friction. In the words of sales theorist J. Winnfield, "Well we'd have to be talkin' about one charming motherfuckin' pig".

Say "what" again! I dare you! I double dare you!!

All of that is fair and I appreciate the comment. Hmm.

#1 - I could fork an instance of bash for every command instead of using an internal interpreter. I've shied away from that because it adds an extra layer of complexity, and also on the theory that if you're trying something that hucksh can't run by itself, then you should probably write a stand-alone script anyway. But perhaps as an option it'd help ease the transition or make people more willing to try it.

#3 - A Windows version is possible. Hucksh has worked there in the past. I think my hurdle there is a build error on Windows in one of the libraries I use. It's probably one of those things that when I actually look closely at it, it'll take an hour to fix, but so far I haven't. More fool I, perhaps.

I also shy away from Windows a tiny bit because it's not my primary platform (macOS), or even my secondary platform (Linux), and the Windows file system structure (with drive letters and so on) requires different code.

On the other hand, a Windows client talking to a Linux server would cover a big use-case.

#4 - If enough people buy it, I could hire other people! ;) I realize that doesn't make it any easier to be an early adopter.

#5 - Would you care to speculate on what you think is a fair price?

Also, based on your other comment, I've added a discount code for 90% off (HN-1223).

I don't think there's anything 'unfair' about your price, it's just something I don't want to have to think about yearly for a shell/terminal. One possible workaround is longer term licenses (especially promotionally to early adopters) or a different paid upgrade cadence.

I think the fundamental problem to me is similar to that expressed by some of the other commenters - a vague sense of unease with using a closed source shell/terminal combo. That's much harder to overcome than paying for a shell/terminal combo. Of course, getting reasonably compensated for open source work is an even trickier problem.

> a vague sense of unease with using a closed source shell/terminal combo. That's much harder to overcome than paying for a shell/terminal combo. Of course, getting reasonably compensated for open source work is an even trickier problem.

I actually see them compounding... either one alone isn't great, but combine them and you're left with a tool that can be rug-pulled out from under you at any time for any reason. OSS with some monetization gives you a fallback of maintaining it yourself, while commercial non-OSS with a perpetual license at least means you can keep using an old version as long as you can manage. Commercial subscription software is at best ephemeral, and not something I'd invest (time, money, transition effort, etc...) in without a strong fallback option, and even then the benefit offered has to be pretty impressive to overcome those switching costs.

> commercial non-OSS with a perpetual license at least means you can keep using an old version as long as you can manage

That's what hucksh has. Today's license will work with today's code (and code released in the next year) forever. Sorry if that was unclear.

Oh, ok, that's at least more reasonable. Thanks for clarifying. (and for using such a license)

Not really unclear, now that I look closer, I just saw the annual rate and hadn't actually clicked through to the license/pricing details and assumed that to mean a typical SaaS subscription model. (it's cool and all, but not a fit for me based on the direction you're going.)

UNC paths are pretty universally supported on modern windows and make things more similar to other platforms. (essentially there are symlinks that make the drive letter look like just an element of the larger path)

Wait, is it a perpetual license with 1 year of upgrades or is it a license that expires after a year? The latter would make this a no-go for me even if I really like everything else that I'm seeing.

It's the former. A license today will work with today's code forever.

I am not a fan of subscription software, and as much as I would love a recurring income, I don't want to inflict that on my users, if I can avoid it. At the same time, I think it's fair to try to have some way to fund ongoing improvements. It's a delicate balancing act.

Thanks, I've bought a license now.

Another question: Is there a publicly accessible changelog anywhere? All I see is a Discord link and a newsletter, neither of which I'm interested in using just to see what's changed in a new version. This is the kind of thing that should have its own page on your Notion and be available as a separate text file in the downloaded zip/tgz.

> Thanks, I've bought a license now.

Thanks, that's awesome!

> Is there a publicly accessible changelog anywhere?

Not at present. I'll definitely do one for future releases. I added a reminder in my Makefile, too :) . Thanks for pointing that out.

Reply on news.ycombinator.com