Comments on: Yegge, Clojure, Arc, and Lolita: or Days of Future Past http://blog.fogus.me/2009/02/06/yegge-clojure-arc-and-lolita-or-days-of-future-past/ λ λ λ Wed, 10 Mar 2010 20:55:55 +0000 http://wordpress.org/?v=2.9.1 hourly 1 By: Jochen Schmidt http://blog.fogus.me/2009/02/06/yegge-clojure-arc-and-lolita-or-days-of-future-past/comment-page-1/#comment-3103 Jochen Schmidt Thu, 12 Feb 2009 10:12:20 +0000 http://blog.fogus.me/?p=1638#comment-3103 <p>Yegge's essay was really fun to read but full of factual errors that showed that he cannot be so experienced in lisp like he likes to tell people. Most of his arguments seemed to be more like hearsay and not like first class experience. I think he is a Paul Graham victim ;-). I used to like Yegge's essays but after "Lisp is not an acceptable Lisp" and his relativizing follow-ups he lost some levels of credibility in my books.</p> <p>What Clojure really proved is that the Common Lisp community is not so religious about other dialects of Lisp like many people thought. Many lispers tolled respect to the design and like to play with it; even use it in projects. Most older "new lisps" seemed to solve mini-problems like unusual naming or were only incompatible because making them compatible would be actual work - it was never clear for more experienced lispers what they actually tried to solve. Clojure shows that it is well designed, solves some real problems and offers new ideas - to me it is no wonder why Lispers like Clojure.</p> <p>(regarding expanding macros) In the LispWorks IDE (Common Lisp) you can macroexpand (expand once) and walk (expand completely) directly in the editor or listener (by key chord or context menu). LispWorks also has a very slick Stepper-Tool. You can place break-points just in your source code (at expressions, not only lines!) and the stepper is automatically opened if the program hits it. The stepper asks if it should automatically expand a macro if it sees one. The whole IDE is running within the same Common Lisp image and is using LispWorks' own cross platform UI-Toolkit CAPI for the Tools. The editor is programmable like emacs; you can integrate this editor as a text-pane in your applications. The only caveat: It's a commercial offering - as some people cannot get themselves using non-free tools, this could be a problem. There is a free Personal editon though - its restricted in memory and you cannot deploy your application with it - but the IDE works like the full edition.</p> Yegge’s essay was really fun to read but full of factual errors that showed that he cannot be so experienced in lisp like he likes to tell people. Most of his arguments seemed to be more like hearsay and not like first class experience. I think he is a Paul Graham victim ;-). I used to like Yegge’s essays but after “Lisp is not an acceptable Lisp” and his relativizing follow-ups he lost some levels of credibility in my books.

What Clojure really proved is that the Common Lisp community is not so religious about other dialects of Lisp like many people thought. Many lispers tolled respect to the design and like to play with it; even use it in projects. Most older “new lisps” seemed to solve mini-problems like unusual naming or were only incompatible because making them compatible would be actual work – it was never clear for more experienced lispers what they actually tried to solve. Clojure shows that it is well designed, solves some real problems and offers new ideas – to me it is no wonder why Lispers like Clojure.

(regarding expanding macros) In the LispWorks IDE (Common Lisp) you can macroexpand (expand once) and walk (expand completely) directly in the editor or listener (by key chord or context menu). LispWorks also has a very slick Stepper-Tool. You can place break-points just in your source code (at expressions, not only lines!) and the stepper is automatically opened if the program hits it. The stepper asks if it should automatically expand a macro if it sees one. The whole IDE is running within the same Common Lisp image and is using LispWorks’ own cross platform UI-Toolkit CAPI for the Tools. The editor is programmable like emacs; you can integrate this editor as a text-pane in your applications. The only caveat: It’s a commercial offering – as some people cannot get themselves using non-free tools, this could be a problem. There is a free Personal editon though – its restricted in memory and you cannot deploy your application with it – but the IDE works like the full edition.

]]> By: Xah Lee http://blog.fogus.me/2009/02/06/yegge-clojure-arc-and-lolita-or-days-of-future-past/comment-page-1/#comment-3102 Xah Lee Thu, 12 Feb 2009 07:12:40 +0000 http://blog.fogus.me/?p=1638#comment-3102 <p>you wrote: «Now I’m not a LISP expert, but it seems to me that the S-Expression is generally regarded as the canonical representation of the LISP AST (and many other languages for that matter). That is, the LISP syntax is as close to an AST as one can get. Perhaps the presence of macros, #, `, ‘, and , muddy the waters a bit, but not much.»</p> <p>that is technically not true. Although you mentioned lisp syntax irregularities like “# , ,@ '”, but in general this sentiment that lisp has regular syntax, or that its syntax is closest to abstract syntax tree, is perpetuating a myth.</p> <p>Mathematica language's syntax, XML, and XML derived general purpose languages (e.g. O:XML, X#), are in fact more regular or closest to a pure tree representation.</p> <p>For detail, see:</p> <p>• Fundamental Problems of Lisp http://xahlee.org/UnixResource_dir/writ/lisp_problems.html (the first section; on syntax irregularity)</p> <p>• The Concepts and Confusions of Prefix, Infix, Postfix and Fully Nested Notations http://xahlee.org/UnixResource_dir/writ/notations.html</p> <p>also note, some pure form of nesting is not the only way to represent a tree. Indendation, such as used in Python or ascii-picture tree used to represent directories and sub dirs and files, is another form, arguably has more clarity than nesting with delimiters. As far as i known, there were a couple such proposal/spec/library with working implementation of this syntax that works for Scheme lisp.</p> you wrote: «Now I’m not a LISP expert, but it seems to me that the S-Expression is generally regarded as the canonical representation of the LISP AST (and many other languages for that matter). That is, the LISP syntax is as close to an AST as one can get. Perhaps the presence of macros, #, `, ‘, and , muddy the waters a bit, but not much.»

that is technically not true. Although you mentioned lisp syntax irregularities like “# , ,@ ‘”, but in general this sentiment that lisp has regular syntax, or that its syntax is closest to abstract syntax tree, is perpetuating a myth.

Mathematica language’s syntax, XML, and XML derived general purpose languages (e.g. O:XML, X#), are in fact more regular or closest to a pure tree representation.

For detail, see:

• Fundamental Problems of Lisp http://xahlee.org/UnixResource_dir/writ/lisp_problems.html (the first section; on syntax irregularity)

• The Concepts and Confusions of Prefix, Infix, Postfix and Fully Nested Notations http://xahlee.org/UnixResource_dir/writ/notations.html

also note, some pure form of nesting is not the only way to represent a tree. Indendation, such as used in Python or ascii-picture tree used to represent directories and sub dirs and files, is another form, arguably has more clarity than nesting with delimiters. As far as i known, there were a couple such proposal/spec/library with working implementation of this syntax that works for Scheme lisp.

]]> By: Austin King http://blog.fogus.me/2009/02/06/yegge-clojure-arc-and-lolita-or-days-of-future-past/comment-page-1/#comment-3101 Austin King Thu, 12 Feb 2009 02:48:06 +0000 http://blog.fogus.me/?p=1638#comment-3101 <p>DrScheme (PLT Scheme) will also Graphically expand your macros.</p> DrScheme (PLT Scheme) will also Graphically expand your macros.

]]> By: rzezeski http://blog.fogus.me/2009/02/06/yegge-clojure-arc-and-lolita-or-days-of-future-past/comment-page-1/#comment-3092 rzezeski Tue, 10 Feb 2009 19:50:47 +0000 http://blog.fogus.me/?p=1638#comment-3092 <p>"It is “Lisp”, not “LISP”. I hate to be that guy, but enough people have been talking about it for long enough to know better."</p> <p>I guess Steele and Sussman should republish the Lambda papers.</p> <p>I for one am loving Clojure. It's made programming fun all over again.</p> “It is “Lisp”, not “LISP”. I hate to be that guy, but enough people have been talking about it for long enough to know better.”

I guess Steele and Sussman should republish the Lambda papers.

I for one am loving Clojure. It’s made programming fun all over again.

]]> By: fogus http://blog.fogus.me/2009/02/06/yegge-clojure-arc-and-lolita-or-days-of-future-past/comment-page-1/#comment-3089 fogus Mon, 09 Feb 2009 17:25:10 +0000 http://blog.fogus.me/?p=1638#comment-3089 <blockquote>It is “Lisp”, not “LISP”. I hate to be that guy, but enough people have been talking about it for long enough to know better.</blockquote> <p>There are two reasons why I like to use LISP instead of Lisp:</p> <ol> <li>I use LISP for the general LISP-ness, and Lisp for a specific Lisp implementation (see how I used it for Common Lisp</li> <li>Using LISP helps to quickly identify the type of people who are usually not interested in having serious discussions about LISP, Lisp, LIsp, lisP, or LIsP.</li> </ol> <p>-m</p> It is “Lisp”, not “LISP”. I hate to be that guy, but enough people have been talking about it for long enough to know better.

There are two reasons why I like to use LISP instead of Lisp:

  1. I use LISP for the general LISP-ness, and Lisp for a specific Lisp implementation (see how I used it for Common Lisp
  2. Using LISP helps to quickly identify the type of people who are usually not interested in having serious discussions about LISP, Lisp, LIsp, lisP, or LIsP.

-m

]]>