On Fri, 2011-10-28 at 09:54 -0700, Mark Engelberg wrote: > On Thu, Oct 27, 2011 at 10:58 AM, daly <d...@axiom-developer.org> wrote: > >> Have there been any new developments on the literate programming / > >> Clojure front, in terms of tools that leverage existing build tools, > >> test suites, generating meaningful line numbers for stack traces and > >> debugging? > >> > > > > The question is ill-posed in sense that literate programming is not > > a tool or technology but a change in mindset. To quote Knuth: > > I think my question is legitimate. > > To take an extreme example, if you tell me literate programming is > great, but that I have to write out my programs by hand on paper and > scan them into a computer, I'd laugh at you. Why would I want to give > up the convenience of typing in my programs via keyboard to do > literate programming?
I'm not sure what tools you use for development. Straight emacs is sufficient for my development. Emacs doesn't care what style is used for development. > > So I'd like an honest appraisal of the state of the art in literate > programming of Clojure -- what tools and conveniences do I have to > give up to do LP. Do I have to give up things like meaningful line > numbers when I get an error message? > State of the art in literate programming in Clojure? Hmmm. The pamphlet link is my only example of literate programming related to Clojure. In order to develop you edit the pamphlet, type "make" at a command line, and the world is rebuilt. Since the code is extracted into the same source tree as it originally had, the line numbers will be related to the file and line as before. My development style involves changing about 10 lines of code or less followed by a complete system rebuild and test cycle. Overlapping rebuild and test is time for writing the literate explanations. Almost always, when something fails it is perfectly clear what lines caused the error. When that isn't the case then a little "navel debugging" (contemplating what might be wrong) is used. I don't use IDEs but I suppose it would be possible to write a literate plugin of some kind that ignored things that were not in code chunks. Line numbers could then be interpreted with respect to the start of the chunk. Most IDEs allow plugins. To your point, though, it is true that we may have reached a local optimum of tool development for tiny files. It might be necessary to move away from this peak in order to climb a different hill. This would imply making things "worse" for development until the tools adapt. Clojure moved off the local optimum of common lisp and broke all my code-walking software, violated quite a few of my assumptions, and invalidated my manuals but that isn't an argument for abandoning Clojure. You might feel that the current toolset is optimal and that literate programming is not worth the time and effort. De gustibus non disputandum. (There is no disputing taste). A literate Clojure would only be of benefit to those who come after. All it offers now is a lot of time and effort to communicate ideas to people we will never meet. However the thing that draws people to Clojure is the ideas, not the code. We can debate it at length over beers at the Conj. Tim Daly -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en
