Provide a better readline experience for Clojure languages.
It is valid to ask why would Clojure programmers need a better readline experience, as our tools of choice normally allow us to edit in the environment of our choice and send off expressions to be evaluated in a REPL.
While this is true, it doesn't hurt to make the REPL experience more helpful and pleasurable when we are forced to interact with it directly at the terminal. These rare situations do pop up, and when they do the anomalous situation is sometimes associated with some kind of urgency. In cases where we are trying to debug a live system, and our tools aren't available, there is no need to make the situation harder with a spartan terminal UX.
So when you do have to use the REPL, it would be nice for it to work better.
However, the main reason for this library is for the newcomers to
Clojure. The path for newcomers to create an effective Clojure
programming environment, is varied, difficult and confusing. It
requires a level of investment and discernment that is too high for
the language explorer who has never used a LISP before. As a result
when newcomers come to Clojure the most intelligent decision they can
make, is to not try and negotiate the tooling needed for an editor
REPL connection but rather just use the
lein repl and/or a edit a file with a familiar editor and constantly
re-run or re-load a script. Thus, they experience a stunted workflow
that is all too familiar in other languages and it is easy to
miss-construe this as the Clojure development experience.
A fluid interactive programming workflow is a fundamental difference that Clojure offers, yet many newcomers will often never see or experience it.
When I refer to a fluid interactive programming workflow in Clojure I am thinking mainly of inline-eval. Most programmers do not have an experience of what inline-eval is. They have nothing to compare it to. They have not used LISPs and SEXPs. You can describe inline eval and demonstrate it to them until you're blue in the face and they won't get it.
It is not until a programmer actually experiences inline-eval as a programming tool that the light goes on. SEXPs start to make more sense, and the why of LISP starts to dawn.
The idea here is to provide the opportunity to experience inline-eval at the first REPL a newcomer tries. The idea is to provide a tool that is sharp enough for newcomers to elegantly solve 4Clojure problems and participate in Advent Of Code without having to make a steep investment in an unfamiliar toolchain.
As a bonus, when we provide this experience at the very first REPL, newcomers will have a base of experience from which they can now draw from to choose their tooling.
They will understand the availability of online docs, source code, and apropos. They will understand the capabilities of inline eval and structural editing. IMHO this experience needs to be communicated as urgently as the other Clojure features.
keep dependency tree very simple and shallow. In order for a library like this to be adopted widely across the Clojure tooling system it needs to not bring extraneous dependencies. This means the core library should have as few dependencies as possible. When dependencies are required the transient dependencies again should be few to none.
JLine is not a small or simple dependency but it is non-negotiable at the moment b/c manipulating a terminal and its capabilities in a cross platform compatible way is a very difficult problem.
provide an exceptional readline experience for Clojure programmers This experience should transcend currently available readline offerings in other languages. Clojure has many advantages (sexps, etc) that provide a readline library a great deal of leverage to do text manipulation.
as a readline library it shouldn't interfere with the input stream when it is not reading a line
as a readline library it is not responsible for REPL output, this doesn't mean it can't provide useful utilities that would help the library consumer, for example to query the last line read, or to redisplay the last line (possibly in place) with a code pointer indicating an error.
open and customizable, following in the example of Emacs this library's behavior should be customizable. Behavior should be modifiable/programable from the readline itself.
This will allow the people to opt in to additional functionality like paredit while keeping the core as simple as possible.
timeline priority is to get something useful into the hands of programmers sooner than later
an eventual goal is to abstract the api enough so that it will work in on a JS platform, this is not an immediate goal.