Articles on the value of strong typing
Hello list, I am a student at the IT University of Copenhagen doing my final thesis on Domain Specific Languages implemented as embedded languages in Ruby. As you might know Ruby uses a somewhat lax form of typing compared to Haskell which is apparently highly utilized for creating embedded DSL's, at least in the academic world. Additionally many of the articles on embedded languages in Haskell use the typing features of Haskell as an argument for using Haskell in this setting. This lead me to the question: Are there any scientific empirical studies of the values of static / stronger type systems as found in Haskell, C# or Java in real world settings? Or any studies comparing weaker type systems in terms of programmer efficiency, defect ratio, etc. I have been unable to find any such studies as it supposedly is "common knowledge" that strong type systems are of great value. Additionally the people I have asked the question was not aware of any such studies, so now I'm trying this list. Please note that I am looking for papers or articles, preferably from the academic community, and not personal opinions (as valid as they may be). -- Thank you, - Jacob Atzen
You may find the following comment interesting. "The mythos of type systems is that they help the programmer. But the reality is compiler and hardware design. Not simply that a fantasy type system is harder to implement, but that a restricted language is easier to implement." page 189 "Theoretical Introduction to Programming" Bruce Mills Springer 2006 I don't know if he bases his conclusions on a study or not.
I have been unable to find any such studies as it supposedly is "common knowledge" that strong type systems are of great value. Additionally the people I have asked the question was not aware of any such studies, so now I'm trying this list.
Please note that I am looking for papers or articles, preferably from the academic community, and not personal opinions (as valid as they may be).
-- Regards, Casey
Casey Hawthorne wrote:
You may find the following comment interesting.
"The mythos of type systems is that they help the programmer. But the reality is compiler and hardware design. Not simply that a fantasy type system is harder to implement, but that a restricted language is easier to implement."
page 189 "Theoretical Introduction to Programming" Bruce Mills Springer 2006
I don't know if he bases his conclusions on a study or not.
Is this section about performance? Otherwise, it makes almost no sense.
On 3/26/07, Jacob Atzen <jacob@jacobatzen.dk> wrote:
This lead me to the question: Are there any scientific empirical studies of the values of static / stronger type systems as found in Haskell, C# or Java in real world settings? Or any studies comparing weaker type systems in terms of programmer efficiency, defect ratio, etc.
It's certainly not scientific, but there is a recent entry on Lambda the Ultimate that contains some empirical evidence: http://lambda-the-ultimate.org/node/2147 -- Martin
As far as published studies, I have found many through the Psychology of Programming Interest Group, which has a web site http://www.ppig.org/ and an archived mailing list http://www.mail-archive.com/discuss%40ppig.org/ with a fairly high density of reports on (formal) empirical studies on related issues. Asking the question on their mailing list is likely to be return quite a few references to proper studies. Jacques
On Mon, Mar 26, 2007 at 03:41:10PM +0200, Jacob Atzen wrote:
This lead me to the question: Are there any scientific empirical studies of the values of static / stronger type systems as found in Haskell, C# or Java in real world settings? Or any studies comparing weaker type systems in terms of programmer efficiency, defect ratio, etc.
I have been unable to find any such studies as it supposedly is "common knowledge" that strong type systems are of great value. Additionally the people I have asked the question was not aware of any such studies, so now I'm trying this list.
I still have not been able to find any studies on the subject even after trawling the ppig conference archives. Thank you for your comments and suggestions. -- Regards, - Jacob Atzen
On Apr 3, 2007, at 12:06 AM, Jacob Atzen wrote:
On Mon, Mar 26, 2007 at 03:41:10PM +0200, Jacob Atzen wrote:
This lead me to the question: Are there any scientific empirical studies of the values of static / stronger type systems as found in Haskell, C# or Java in real world settings? Or any studies comparing weaker type systems in terms of programmer efficiency, defect ratio, etc.
I have been unable to find any such studies as it supposedly is "common knowledge" that strong type systems are of great value. Additionally the people I have asked the question was not aware of any such studies, so now I'm trying this list.
I still have not been able to find any studies on the subject even after trawling the ppig conference archives.
The following article might be relevant. @string{TSE = "{IEEE Transactions on Software Engineering}"} @ARTICLE{PT98, author = {Prechelt, L. and Tichy, W.~F.}, title = "{A Controlled Experiment to Assess the Benefits of Procedure Argument Type Checking}", journal = TSE, volume = 24, number = 4, year = 1998, pages = {302--312}, } -- Martin
participants (6)
-
Casey Hawthorne -
Derek Elkins -
Jacob Atzen -
Jacques Carette -
Martin Erwig -
martin odersky