Lisp as an Alternative to Java

Joseph Dale writes "Lisp as an Alternative to Java is a detailed and well-reasoned study comparing Lisp to Java and C++ in terms of execution time, memory consumption, and developer effort. The author, Erann Gat, was the principal software architect for the Mars Science Microrover, the prototype for the Mars Pathfinder rover."

Comments

Re:It's about the API

[lars_stefan_axelsson]

Well, all that repetition that you claim one must do to program in Lisp (Scheme in my case) should translate to longer development times, shouldn't it?

Then how do you explain that the development times reported for Lisp were about half those of Java, comparing the median?

Re:Interesting, but flawed?

[nels_tomlinson]

Some of thing things I believe are more important when thinking about a programming language:

1) Amenable to use by team of programmers
2) Viability over a period of time (5-10 years).
3) Large developer base
4) Cross platform - not because I think cross-platform is a good thing by itself; rather, I think its important to avoid being locked-in to a single hardware or Operating System vendor.
5) Mature IDE, debugging tools, and compilers.
6) Wide applicability

1) Is lisp less amenable to use by a team? Can't you implement one object while I implement another? I'm asking because I don't know; all my programming has been for my own research.

2) Lisp has been around since the 1950's, in one form or another. I find that I can read lisp code from the Lisp 1.5 programmer's manual, which originated in 1965. Our kids will probably still be able to in another 30 years.

3) How large is large enough? Is: "language of choice for 10^7 illiterate script-kiddies" a stronger recommendation than: "there are a sufficient number of competent programmers who are proficient in the language and familiar with the problem domain"? Notice that second statement is probably not true for any of the languages discussed, for most problem domains.

4) Except for the Microsoft offerings, pretty much everything is cross platform. Think gcc and cmucl, for easy examples.

5) I really don't know what an IDE is. I gather it is something other than emacs, since emacs supports nearly every language, some wonderfully well (try ESS for R). I really can't comment on this one at all; perhaps someone else could?
As for mature compilers, again, I think all of the languages mentioned so far are in fair shape there. Java and c can both use the same back end in gcc, and so does gcl.

6) Technically, all the languages mentioned are Turing-complete ... But wide applicability means being suitable, instead of merely capable. I am not familiar with java and c++, but lisp and c (and Fortran) seem to be equally capable for the bits of number crunching I do. Lisp is wonderfully more flexible, and the debug/wish-i'd-done-it-differently/change it/debug/repeat cycle seems a good deal shorter in Lisp than the other two, for me at least.

After going over the above, it looks to me as if personal choice and the resources immediately at hand are really the only reasons to pick lisp over c over c++ over java over lisp.

Re:What about GUIs?

[prizog]

Java's event-driven GUI system is easy to code for, but is limited by the simplistic event model. Still, it does not make much totally impossible, so in my book, it is decent.

I am not sure what you mean by a packing layout manager (but you can tell me next time you see me on IRC).

I do see a list of toolkits, but they all seem to be based on terrible base toolkits, like Tk or WxWindows, or are not portable at all.

CLIM is a good idea, but I cannot find any free (beer) implementations, which seriously limits its usefulness.

Swing has the property that, since it is written almost entirely in Java, it can be ported with little effort. Bindings need not be written for each component, but only for windows (Xlib meaning) and a few other things. Unfortunately, this tends to make it a bit slow, but that's not a concern for me.

Re:I'm a professional who uses Java

[sigue]

Right on the money, honey. I work at a company that has an e-commerce web site written in Common Lisp. When we got VC funding one of the requirements was that we eventually rewrite the whole thing in Java. Not for technical reasons, but because of their fear that we wouldn't be able to find enough good Lisp programmers. Well, we hired a boatload of Java programmers. They got approximately nothing done in about a year. We four Lisp hackers kept the site from grinding to a halt while they spent our money on Java dreams. VCs pulled out, we got rid of the Java programmers (and marketing spend-thrifts), we decided to continue with the conversion to Java, we generated 50K lines of Java code from the Lisp introspection facilities, we found that we had 10x code bloat in Java, dumped java, yay, back where we started, actually improving the site now!

One striking thing was the complete lack of interest most of the Java programmers seemed to have in learning anything about the Lisp code. I guess they thought "how is this going to help my career"?

Lisp is perfectly suited to web applications. Since it's mostly just application logic and sending bits out the wire, Lisp's ability to change code on the fly without restarting app servers and such is a huge win, especially since a web site often requires frequent, small updates.