This theoretical hope doesn't pan out in postmodern practice ; as it turns out out, of such pharmaceutical abundance as besets us on all sides some drug can always be found or tailored for every hurt boi, so that ever-precious neoteny may be preserved.

Otherwise, this "better tool" thing is a fool's errand. There are "available" in the very limited sense here discussed uncounted (because uncountable) multitudes of such non-tools, bees carved out of birds, maimed tanks masquerading as machine guns (always inadequately deployed, too, for bonus horror). They are not things. A "scheme" replacement for bash is not a thing, not anymore than any other stupid idea of rambunctious boyhood, the socks with hoodies and their ilk. You are welcome to repeat the experience as many times as you can carry, but it'll always come out to the same way : doing just short of enough in evidently too much space, with obviously too much weight, by clearly too much money. Hitlerism, in a word, something which by well documented experience can sink organisations also, and just as well.

Well, if you look at all the things that were historically proposed as "replacements", all of them are dead now (I guess with the exception of fish shell, which is only a minor departure from bash). In this sense, I personally would not attempt such project personally. And the problem is not only too much space, the hypothetical replacement should have enough stability in time as well (no breaking of syntax/features), which combined with very large space to cover makes writing a hypothetical replacement an ~impossible task. Perhaps, the time for providing better shell replacement was in mid-80ties, and there is no way out now.

That said, IMO it's not like there is a usable Scheme implementation either (most of them are either overcomplicated or too limited) - so I still won't be stopping anyone from doing a Scheme in Ada; and having access to shell utilities (and I would add direct access to Linux syscalls as well) is the only way it could conceivably used for anything useful. Of course, most likely the point, that in a year or so, those who started to use it for long-term things would curse rewriting everything back to bash, will stand.

As to the shit-and-sticks, the point was made (and the attempt to bury such toxic facts followed immediately). Nevertheless, the point stands : A + epsilon is not a simplification over A alone, even if "I could just dwell in the epsilon my whole life and pretend A doesn't exist" is a solution "that seems to work okay" for... again... retarded bois. Moreover, secondarily but not to be entirely omitted : those "shit and sticks" work way the fuck better (in actual, lived practice) than anything else "available" (which it isn't, even though it may "convincingly" pretend to be). Always "fixing" the parts that didn't need any fixing at all is how the useless bois club stays useless over time, how its shit is always "inexplicably" late and so on. Forever, or at least that'd be their hope.

This is a nice reference for this article, not that I can somehow argue against it. And yes, when something needs to be done, it's those utilities that are used, irrespective of their potential warts.

Which is also why, (this goes to @Stanislav) I never liked esthlos' CL-V: why reimplement Keccak in CL when there is perfectly working one in Ada, available over pipe, and in this case - without warts at all? This is why IMO implementing signature verifier in CL/Scheme is wrong approach (yes, I know that if you parse the signature, verification is a ~3 lines of code, if the Scheme has actual bignums): you can use GPG, or perhaps Peh as well, why duplicate the work? And then rewrite phf's vpatch in CL as well, for the argument of purism?

Otherwise, this "better tool" thing is a fool's errand. There are "available" in the very limited sense here discussed uncounted (because uncountable) multitudes of such non-tools, bees carved out of birds, maimed tanks masquerading as machine guns (always inadequately deployed, too, for bonus horror). They are not things. A "scheme" replacement for bash is not a thing, not anymore than any other stupid idea of rambunctious boyhood, the socks with hoodies and their ilk. You are welcome to repeat the experience as many times as you can carry, but it'll always come out to the same way : doing just short of enough in evidently too much space, with obviously too much weight, by clearly too much money. Hitlerism, in a word, something which by well documented experience can sink organisations also, and just as well.

As to the shit-and-sticks, the point was made (and the attempt to bury such toxic facts followed immediately). Nevertheless, the point stands : A + epsilon is not a simplification over A alone, even if "I could just dwell in the epsilon my whole life and pretend A doesn't exist" is a solution "that seems to work okay" for... again... retarded bois. Moreover, secondarily but not to be entirely omitted : those "shit and sticks" work way the fuck better (in actual, lived practice) than anything else "available" (which it isn't, even though it may "convincingly" pretend to be). Always "fixing" the parts that didn't need any fixing at all is how the useless bois club stays useless over time, how its shit is always "inexplicably" late and so on. Forever, or at least that'd be their hope.

Theoretically -- would not need pipe, if a native (i.e. in Scheme) sig verifier were added.

@Diana:
Fully agree with both statements.

The thing with standardization is that there has to be someone owning it - aka in practical terms it can only be done by an authority. Lacking that, it's pretty much a matter of "better stick to what shit you are already stuck with, anyway" indeed.

Seems like there is no hash verification though (just as there wasn't in my orig. pythonistic prototype Vtron) ?