Comments on: The Definitive Programming Language, Not http://www.lshift.net/blog/2006/06/13/the-definitive-programming-language-not What happens at LShift Wed, 07 Jan 2009 02:36:55 +0000 http://wordpress.org/?v=2.0.12-alpha by: Peter Van Roy http://www.lshift.net/blog/2006/06/13/the-definitive-programming-language-not#comment-65112 Sat, 01 Sep 2007 10:05:13 +0000 http://www.lshift.net/blog/2006/06/13/the-definitive-programming-language-not#comment-65112 <p>Unfortunately, the critique is flawed. Its conclusion is preordained and the evidence is carefully selected and trimmed to support its conclusion.</p> <p>For example:</p> <p><i>the main driver for Oz - distributed programming</i>: this is false. Oz was essentially complete when the work on distribution started in 1995.</p> <p><i>Language design follows trends.</i>: Joe Armstrong (Erlang designer), Mark Miller (E designer), Gert Smolka and his colleagues (Oz designer) are independent thinkers that are not so much influenced by trends. They all worked <em>against</em> the trend of their surroundings, in independent fashion.</p> <p><i>This implies distributed programming.</i> Reducing everything to the needs of distributed programming is just not what happened. For example, Erlang was originally not designed as a distributed language, but as a language that allows programs to survive software faults.</p> <p><i>There are not four languages.</i> The paper clearly explains that there are three languages and in how far that this can affect the conclusion. This is not an original observation in the critique.</p> <p><i>Unfortunately these claims do not stand up to scrutiny.</i> The scrutiny is itself fundamentally flawed.</p> Unfortunately, the critique is flawed. Its conclusion is preordained and the evidence is carefully selected and trimmed to support its conclusion.

For example:

the main driver for Oz - distributed programming: this is false. Oz was essentially complete when the work on distribution started in 1995.

Language design follows trends.: Joe Armstrong (Erlang designer), Mark Miller (E designer), Gert Smolka and his colleagues (Oz designer) are independent thinkers that are not so much influenced by trends. They all worked against the trend of their surroundings, in independent fashion.

This implies distributed programming. Reducing everything to the needs of distributed programming is just not what happened. For example, Erlang was originally not designed as a distributed language, but as a language that allows programs to survive software faults.

There are not four languages. The paper clearly explains that there are three languages and in how far that this can affect the conclusion. This is not an original observation in the critique.

Unfortunately these claims do not stand up to scrutiny. The scrutiny is itself fundamentally flawed.

]]> by: Robert http://www.lshift.net/blog/2006/06/13/the-definitive-programming-language-not#comment-24331 Mon, 18 Dec 2006 01:49:16 +0000 http://www.lshift.net/blog/2006/06/13/the-definitive-programming-language-not#comment-24331 <p>There are all-purpose programming languages and there are specialized programming languages. Aside from SQL, specialized programming languages do not get enough respect from mainstream techies, as I'll argue in an upcoming blog entry. </p> <p>Take statistical programming languages: the three most well-known are S, SAS, and SPSS. They are all at least 30 years old. We have not seen a ground-breaking statistical programming language break out into the mainstream for more than three decades. </p> <p>There is a new alternative, VILNO, available as an attachment to the August 31 blog entry at www.my.opera.com/datahelper</p> <p>Robert</p> There are all-purpose programming languages and there are specialized programming languages. Aside from SQL, specialized programming languages do not get enough respect from mainstream techies, as I’ll argue in an upcoming blog entry.

Take statistical programming languages: the three most well-known are S, SAS, and SPSS. They are all at least 30 years old. We have not seen a ground-breaking statistical programming language break out into the mainstream for more than three decades.

There is a new alternative, VILNO, available as an attachment to the August 31 blog entry at
www.my.opera.com/datahelper

Robert

]]> by: Philip Wadler http://www.lshift.net/blog/2006/06/13/the-definitive-programming-language-not#comment-3308 Wed, 14 Jun 2006 09:12:59 +0000 http://www.lshift.net/blog/2006/06/13/the-definitive-programming-language-not#comment-3308 <p>This paper was invited as a talk at FLOPS (I was responsible for this, as pc chair). Rather than focussing just on his own work, on Oz, Peter van Roy chose to also publicize the work on Erlang and E. Researchers spend too little of our time extolling the virtues of others work, so I was delighted to see this.</p> <p>I agree with yoiu about the shortcomings of the analysis, but I also agree with you that the paper has a lot of merit.</p> This paper was invited as a talk at FLOPS (I was responsible for this, as pc chair). Rather than focussing just on his own work, on Oz, Peter van Roy chose to also publicize the work on Erlang and E. Researchers spend too little of our time extolling the virtues of others work, so I was delighted to see this.

I agree with yoiu about the shortcomings of the analysis, but I also agree with you that the paper has a lot of merit.

]]>