Hacker News Reader

Comment by yladiz | Hacker News Reader

I would say the separation of concerns is very optimal right now. The terminal emulates the shell, and that's it; there's nothing going on in between or in the terminal besides displaying what the shell outputs. I don't really need mixing of the two, which this looks like it does. For example, if I entered a URL into the terminal, it should be up to the shell to interpret it, not the terminal.

This is not to say that I don't mind "magic" -- I removed the `git` prefix from my git commands because I like typing `add .` rather than `git add .` all the time -- but I don't want my terminal to do it, I want my shell to.

View on HN

Is customising the shell prompt with a PS1 variable good though? The result often lacks clarity.

I don't think the shell should be handling styling.

I also think the shell became less useful when modern application development began to happen on the Web. More should be done to make it easier to interact with Web applications from the command line.

Sorry but webapps make up a tiny portion of what I need to interact with on the CLI as an admin. When I do need to interact with webapps I use an API using curl or python (depending what I need to do).

When do you need to interact heavily with a webapp? Maybe I'm missing something here.

For example, say I ran a command to show disk space usage. Perhaps I would like to pipe the output into a d3 visualisation and see it directly within the terminal without context switching. Why shouldn't I be allowed to do this?

Have you read "The Visual Display of Quantitative Information" by Edward Tufte? Text isn't the best representation of all types of data.

Imagine also that I wanted to represent a table of data within an Excel style spreadsheet. This could be useful as it could define affordances for a user such as the ability to select rows and columns, and to sort or filter in realtime.

I can see the value in that, I don't see how looking across to a browser which has rendered the results of a command piped into a script is the end of the world.

One issue I have with bluring the lines is you open yourself up to a huge attack vector. If I can craft the output of my shell command to get code execution within your terminal I can now do anything, anything on any system you have access to.

Yes - it is a possible attack vector but that doesn't necessarily mean it shouldn't be attempted.

Reply on news.ycombinator.com