Re: Dylan 2016: Evolving Dylan
Răzvan Rotaru <[email protected]> Fri, 25 Apr 2014 15:26:46 +0300
| Newsgroups | gmane.comp.lang.dylan.gwydion.devel |
|---|---|
| Message-ID | <CAP33cnZMz=RUbpxnZyvOxVFvLeCOzj1xu3pbxqXAc-z4G44fKQ@mail.gmail.com> |
--===============0615168863== Content-Type: multipart/alternative; boundary=001a11c2c3383ef38b04f7dd1403 --001a11c2c3383ef38b04f7dd1403 Content-Type: text/plain; charset=UTF-8 Hi, I'm new to Dylan so I don't really know the principles on which it is based or the design decisions that it made. But if I were to think about building or evolving a language I would definitely focus on at least one concurrency strategy. I would want to adapt to the hardware available today, and this is multi-core. I would also focus on immutability. In his talk "The Value of Values", Rich Hickey completely convinced me how important immutability is and how I should be looking for it, whatever I do. The solution doesn't have to be the same as Clojure's (immutable and persistent core data structures), although I don't know how else you could approach it. But what I would want is to have language support for immutability and a tendency towards it. I am thinking that maybe classes can have structural sharing by simply sharing references to the values of its slots, without the use of any trees (in other words each slot would be a reference to the value). Basically the class would become a named collection of references to immutable values. It also wouldn't be very hard to implement, and whatever the performance penalty is, it should be countered by efficient use of available cores. Razvan Rotaru --001a11c2c3383ef38b04f7dd1403 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable <div dir=3D"ltr"><div>Hi,<br><br></div><div>I'm new to Dylan so I don&#= 39;t really know the principles on which it is based or the design decision= s that it made. But if I were to think about building or evolving a languag= e I would definitely focus on at least one concurrency strategy. I would wa= nt to adapt to the hardware available today, and this is multi-core.<br> <br>I would also focus on immutability. In his talk "The Value of Valu= es", Rich Hickey completely convinced me how important immutability is= and how I should be looking for it, whatever I do. The solution doesn'= t have to be the same as Clojure's (immutable and persistent core data = structures), although I don't know how else you could approach it. But = what I would want is to have language support for immutability and a tenden= cy towards it. <br> <br>I am thinking that maybe classes can have structural sharing by simply = sharing references to the values of its slots, without the use of any trees= (in other words each slot would be a reference to the value). Basically th= e class would become a named collection of references to immutable values. = It also wouldn't be very hard to implement, and whatever the performanc= e penalty is, it should be countered by efficient use of available cores.<b= r> <br></div><div>Razvan Rotaru<br></div></div> --001a11c2c3383ef38b04f7dd1403-- --===============0615168863== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ hackers mailing list [email protected] https://lists.opendylan.org/mailman/listinfo/hackers --===============0615168863==--