Hacker News new | ask | show | jobs
by nameless912 3636 days ago
There's literally nothing about this I like.

A terminal with a gradient background might look pretty for screenshots, but for long sessions it just adds to fatigue and makes text harder to read. "Saving your favorite commands" in a normal terminal involves writing an alias into your rc file, not a bunch of button clicks. Same goes for the whole "interactive flags" thing.

Whoever made this clearly hates the terminal and wanted to bring a hipstery web 2.0 approach to the terminal. My question is, WHY?

Oh, and it's PAID? Fuck that. Learn to use a proper terminal or don't bother, simple as that. That isn't me being a crotchety old timer, that's me giving you serious life advice. You will always be faster in the terminal if you learn to use it properly. a 2 hour tutorial in shell scripting and learning to use `man` is worth far more to your career than 10 bucks. Learn it the right way if you're going to bother.

</rant>

3 comments

I have several devs in the office that have to use the terminal a grand total of 2 times a month, and even then I installed some shortcuts for them to make using adb easier.

Being able to build a simple interface on top of the few commands they use, so they don't have to remember flags, argument orders, etc., would be advantageous to us all (and require less support on my part).

Also, what is possibly wrong with someone trying to make some money for their work??

There's nothing inherently wrong with wanting to make money for your work. The problem, honestly, is that a good terminal is a solved problem (c.f. urxvt, which is perfect in every way). Charging people for one would be like charging people for a vim clone. Why charge for an inferior product when someone's already created perfection? ;)

Second of all, learning shell scripting is an almost ridiculously important skill. Wouldn't it be better that you sit those devs down and have them go through a couple hours of intensive shell boot camp so they don't need your help anymore? Anyone who's a halfway-decent dev can learn the basics of shell (simple scripting, variables, redirection, piping) within a couple of hours, easy.

To say the problem is solved you must first state it.

Maybe your terminal problem is solved (mine is), but their problem may be different.

Would you tell people who made chrome to quit back then, because the browser problem was solved? (IE, Opera, Mozilla Firefox) Or maybe tell Google to abandon drive and Microsoft onedrive and all that smaller companies like SugarSync to drop it because Dropbox has "solved it"?

This also means that indeed your previous question is good: it'd be nice if the author could explain the use case and business case for his terminal, seeing that we have a number of others around. (Google explained why Chrome with a comics thingy and got me immediately). But your questioning sounded like the author was already in err.

It's not that simple. The shell for you and me is muscle memory. For someone who takes a "boot camp" and then doesn't use what they've learned for 3 weeks, the boot camp was a waste of time.

Haven't you ever come back to a piece of code from a month ago and had to reacquaint yourself with it? (I know I can't be unique in that... ;) ) We cram an awful lot into our heads as devs, there's no room for things we barely use from a month ago.

To you, the shell may be perfect. For someone else, it's a severely antiquated method of input that was outdated after Windows 3.1.

I love me some shell, but that's you and me. Not all devs.

Edit: reminds me of CPR class. After a month or so, you better brush up lest you actually need to use it!

Here's what you do: in the middle of the night, sneak into all of your devs' offices and fuck with their xorg.conf's. Make them drop to command line only for a couple weeks (because we all know it takes at least that long to diagnose an xorg.conf....) and they'll learn. Oh they'll learn ;)
That's like saying you drop a modern human in the middle of the jungle, and tell em to survive for a week. I'm sure they'll "learn", or die trying but is that really a good way to do it?!
Or they'll quit. Oh they'll quit ;)
or, they'll just reinstall the OS, if they have decent backups.
you're a terrible person
I also kick puppies and write one word commit messages.
But I like them.
While I agree with you that urxvt is a great tool, I disagree with the suggestion that the terminal is "a solved problem" (or that vim is perfection). There is a lot of room for innovation, though not necessarily in the direction that Go Terminal is going.
I agree with each and every one of your comments; in addition, I would like to know how this project was created: what language it was made in, why the author found it necessary to create it, etc.
> what language it was made in

HTML + JavaScript + Electron

Just download that archive, extract its content, and you will find the "goterminal" script that is basically an ASAR archive [1] with a copy of CEF [2] as the webview for the web application that they (for some unknown reason) decided to create.

[1] https://github.com/electron/asar

[2] https://bitbucket.org/chromiumembedded/cef

> that they (for some unknown reason) decided to create.

That's unpleasant. This is a forum for programmers. Care to explain why you are criticizing a participant for programming?

I see at least two possible reasons to create such a terminal: One: The author might have problems with how other terminals work. The other: The author just wanted to create something, without providing a superior value proposition. If it is indeed the latter scenario, it just wastes others' time.
> If it is indeed the latter scenario, it just wastes others' time.

What an absurd, and disgusting, point of view. The author is free to create whatever the fuck he/she wants, for fun. It's not wasting anyone's time -- no one else is obliged to spend any time interacting with it in any way.

You should really reconsider your position. One point is that, while you may have the skills and experience to only work on projects that improve the status quo, many of us do not; and even those who do, as beginners had to pass through the stage where they did not.

If a project does not improve the status quo, I think it should not be presented as something one “should have”.

The time wasted is exactly the time it takes to find out that “elegant and efficient” is at best hyperbole, at worst a lie.

> HTML + JavaScript + Electron

flips over table

Actually, let me unflip that table.

I am desperate for someone to come along and realize that the current terminal interface is a perpetual train wreck, and that if we can break compatibility but end up with something better we should.

Consider for a moment what a monumental task it is to make a simple ncurses menu and status bar in bash script. Consider for a second how trivial that could be made if each command line entry were a small flexible webview.

Imagine if you typed netstat and it showed the results, but there was also a field where you could key in text narrowing queries as part of the standard presentation for all very long input. No more appealing to grep or more and losing context or blasting your scrollback buffer.

Consider for a moment if we could move away from trivial line interfaces for SSH and towards a protocol that actually handles sessions and works over SSL in every conceivable network environment. Hell, most people I know have migrated to mosh and only type ssh when they're provisioning fresh cloud instances (usually to get mosh on there as fast as humanly possible).

The entire idea that "text is a superior interface" is correct as an input mechanism, but not necessarily as a data presentation mechanism. HTML and some basic es2015/typescript/whatever isn't just flat out better than bash for this sort of thing, but it flows from a long history of tools like tcl/tk trying to make the shell a more efficient and effective place to visualize trivial data.

I've been using terminals since 80 characters was a luxury and amber was the fancy new color, and for at least 20 of those years I've been hoping we'll see something that keeps the composability of shell scripting but introduces a non-line oriented protocol that could also offer presentation logic when appropriate.

We could preserve all of that if we just pretend all existing shell commands return a (forgive the templating syntax, it'll be understandable by a wide audience) Map<Stream<Text>> where {'stdout': <stdout>} and additional data can be offered around that with trivial API additions.

Please, let's not pretend we'd not benefit massively from richer tooling here.

> Please, let's not pretend we'd not benefit massively from richer tooling here.

Absolutely nothing about this project or the way they went about building it shows any tendency towards richer tooling. A web interface also does not inherently enable richer tooling. You could implement this non-line oriented protocol entirely without 500+ MB of JS and the interpretation of that data could obviously be done without it too.

Edit:

While I understand your rants all over these threads (it doesn't inherently have to be inefficient and bloated), you've picked a strange submission to take these fights in. The software in question is in fact bloated, not compliant and is really only an interface to aliases.

Also, I would add: The proof is in the pudding. You seem to have made this a personal crusade, so I would urge you to make a decent PoC with these ideas to show the world.

> you've picked a strange submission to take these fights in.

I don't see how that's germane. I didn't object to concrete references to the software (which is not as good as other submissions). I objected to knee-jerk silverbacking.

> You seem to have made this a personal crusade,

No, but I am so tired of people on HN mimicking the style of development we put together in the 80's and 90's because... well I have my opinions about this practice, but they're uncharitable. We'll leave it at that.

There is a cult of terminals in our industry that has fundamentally missed why terminals were an effective interface and throw shade on anything that isn't a terminal, even when it's actually a refinement on the terminal concepts.

> Consider for a moment what a monumental task it is to make a simple ncurses menu and status bar in bash script.

It really isn't, see e.g. whiptail.

And using ncurses for anything today is not a good idea. Surviving terminals are overwhelmingly ANSI, and can be supported by a single codebase that uses escape sequences directly, wihtout the horrors of ncurses.

I have seen whiptail. It's still easier to make a list of selectable options in html or markdown.
In your vision where text is not the universal interface, how would Unix pipes work? After all, often the output of one tool is the input of another – that usually has not been written to interact with the first.
Powershell accomplishes this by making the pipes pass tagged objects around containing streams. This actually has a lot of nice properties when you go to manipulate the objects programatically.

And it's a logical extension of the UNIX stream model, where in reality you're passing around a whole heck of a lot more than text when you type 'cat | sed'. You're passing environment context, multiple streams, etc.

The idea that this is "more complex" than unix pipes is understandable, but the reality is that IPC is pretty complex and a whole lot of magic and shared expectations go into the "simplicity" that bash offers. And even then that model breaks down all the time (e.g., as Docker becomes common as a deploy target the idea that you should output error conditions to STDERR has rapidly become problematic).

How do you imagine scrollback/ history on such thing? Genuinely curious. Have seen a few projects attempting this, but having all of the output searchable is a very central feature of terminal, for me.
Yikes. Thats a great way to make a terminal into a memory hog and unresponsive under load.
Oh?

Historically most terminal implementations have been fraught with performance issues. We'd moved off xterm to aterm and then back. We'd been hesitant to adopt gnome-terminal because of the GTK+ chrome then saw its font kit actually rendered faster.

What most people don't get is how absurdly well-optimized the text rendering and reflowing is for web browsers. With care, people can get modern cellphone browsers to hit 60fps animations and reflows while pushing huge volumes of text through them.

Take any 3 GUI toolkits you can name and ask yourself, "Do all three of these combined have a maintenance budget that equals that of even Firefox by itself, let alone any TWO modern web browsers?" Hell, some of the most cutting edge research into interpreters and garbage collectors is coming out of Google making V8 fast.

Quite frankly, I'd bet that with care and foresight this could be faster than iTerm.

Maybe I am wrong, but I have yet to see a modern browser take anywhere near as little memory of CPU as a modern terminal emulator.

There is also the matter that no matter how much optimization you put into something you can't recoup the cost of the fundamental design choices you went with. Choosing something with less power can solve a lot of headaches.

> little memory of CPU as a modern terminal emulator.

Browser instances can be quite svelte. But uh... I mean... shared lib pages tend to be actually shared pages. That's why the physical memory for multiple browser instances is a lot lower.

What's more, I'd gladly pay a higher price for a richer experience. We no longer need to or should rely on ANSI interfaces. In fact, they're becoming increasingly awkward in a world that really wants endpoints to be stateless by design rather than stateful sockets.

Many people have already discarded the single socket model oh SSH in favor of Mosh. Let's also ditch the console for all but the most basic of tasks. It's not 1997, the lingering stink of proprietary rich client environments is cleansed. Web browsers have annihilated that assumption and made something truly open and pretty amazing.

I use a web browser as my core text editor and I miss emacs not a whit and quite honestly it feels more responsive than my emacs setup. I don't notice any tangible performance differences from Vim in a terminal.

We can do better.

What machine are you running your terminals in? I have as little as 8 GB of RAM, and more CPU power than I could dream a decade ago.

Don't you think that maybe you are using your past environment to judge a current experience?

Visual Studio Code is based on Electron and works wonderfully. Doesn't feel bloated, it's quite fast to open and I haven't felt a slow down in any meaningful way.

I'm not imposing my computer specs to anyone (and I'm using vs code in a client machine in any case), but we have generally a lot more horsepower now and the discussion of performance might mean little for a desktop app (heck, we are running javascript on the server nowadays!).

Oh god, the last software I'd want to be running a terminal in!
>HTML + JavaScript + Electron

This was my creeping suspicion the moment I saw the screenshots. Utter garbage.

I'm cringing at that entire paragraph.
I'm cringing.
> why the author found it necessary to create it

That's unpleasant. This is a forum for programmers. Care to explain why you are criticizing a participant for programming?

You seem to have made this about you, which is fine, but why should we care what _you_ like? Are you a person whose opinion is deeply valued and sought after? Because if you are, then certainly a subjective statement will be weighted by the knowledge, experience and the expertise of the person saying it. So my question is, why does it bother _YOU_ if someone, somewhere, finds value in this and prefers this to whatever other way you think other people should do it?
This isn't Facebook or Reddit. This is HN. Anybody trying to pass that thing off in this community has got some criticism coming.