Hacker News new | comments | show | ask | jobs | submit login

Great article but isn't a lot of this more about using different levels of abstraction i.e. frameworks, DSLs and specific to those? That could either take the form of being sufficiently focused that a simple text editor is adequate (e.g. high level commands, no cruft), or a full-featured managed environment (state, docs etc) specific to the framework. I'm not seeing how this is a task for a generic IDE unless I missed something about light table and they are creating a new language.



> unless I missed something about light table and they are creating a new language.

http://www.lighttable.com/2014/03/27/toward-a-better-program...

The short version is yes, we are making a new language. The kind of workflow I discussed in the OP was the original goal for Light Table and it turns out that it can't just be hacked onto existing languages.


Fair enough. I won't say your decision to make a new language is right or wrong - I don't know how it will turn out - but given that you have decided to do so, and given your objectives for it, my one big recommendation would be to make relational databases central to the language, the way Perl makes string processing central and PHP makes web pages central. Some possible ideas are outlined at http://geocities.com/tablizer/top.htm - you don't have to go with those particular ideas of course, but I think it's important to make relational data central one way or the other.


We are heavily inspired by http://shaffner.us/cs/papers/tarpit.pdf




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact

Search: