I would be more sympathetic except that I am constantly seeing a certain subset of Haskell programmers engaged in the same kind of advocacy. Should they all disclose a conflict of interest because they like Haskell?
If they contributed to the Haskell type system and they are commenting on the inferiority of the type systems of other languages, absolutely yes.
This is far from his first Haskell-bashing exercise. Take a look at the Wikipedia page for Haskell: "Robert Harper, using Standard ML to teach introductory programming, has given his reasons for not using Haskell. Among these are the difficulty of reasoning about resource usage with non-strict evaluation, that laziness complicates the definition of data types and inductive reasoning,[49]and the "inferiority" of Haskell's class system compared to ML's module system.[50]"
I'm not saying his arguments aren't valid. It's just curious that he's not attacking SML with the same vitriol, and perhaps people would like to know why.
I think in Harper's case he really does believe SML does it The Right Way, rather than promoting SML for ulterior/career reasons. His opinion on that subject isn't shared by many, but I think it's a genuinely held opinion. He just has very particular views on how type systems should be, and tends to overstate his views (e.g. treating questions reasonable people can disagree over as if he were stating unarguable facts).
I don't see how he is obligated to attack anything else with the same vitriol, if he doesn't see the same problems with it. It seems that we might just as well try to declare a moratorium on Haskell programmers' bashing of other languages on the basis that those other languages are not type-safe enough.
Any time someone bashes $FAVORITE, I can just disclose to everyone that they are a ${FAVORITE}-basher and therefore aren't speaking credibly. Hooray!
Surely you don't believe that SML doesn't have its own set of issues? My point is he doesn't criticize SML in the same manner on any issue, for a very good reason: he helped design the language.
I believe SML doesn't have any issue that Robert Harper cares about. It is nearly perfect in all things Robert Harper cares about. Are you suggesting that Robert Harper should care about more things than what he currently cares about?
I disagree with your opinion that he doesn't criticize SML because he was involved in the design. SML is not perfect, but Robert Harper does not criticize SML because SML is Robert-Harper-perfect. Or nearly so.
His argument can be valid and still be pointless for 99% of people. Granted we all must make these decisions for ourselves, but if someone smirks and says "screw Haskell" because the exception system is unsound and then goes back to using Perl or Java, they have missed the point.
> I would be more sympathetic except that I am constantly seeing a certain subset of Haskell programmers engaged in the same kind of advocacy. Should they all disclose a conflict of interest because they like Haskell?
If you see someone engage in misleading advocacy you should call them out for it. A big problem with Haskell is that it is so weird and different that new users are always blogging about how great it is. When they make a small realization they blog it, and their misconceptions come along with it. I've been using Haskell for a while, and I find it very hard to stomach it when this happens, so I don't usually follow the comments closely when I see it happen. I doubt I'm alone in that. But there's a world of difference between newbie love babble and a calculated PR stunt by a PL expert.
One can craft an entire argument that's valid and still complete BS. Politicians do it every day. You should be asking if his arguments are sound given adequate context and background information. Harper is most definitely biased against Haskell as evidenced by criticisms of his examples and how he plugs SML. Given this, the facts and examples he provides on Haskell shouldn't be taken at face value. Valid arguments are easy to spot. Sound ones are not.
To be fair, he did write a book on programming languages and he knows a lot about the subject. He might troll a little but IMO it's always enlightened trolling.
In this case, he points out a corner case that the Haskell community already knew about. But I, a beginner haskeller, didn't.
It's a much higher level of discourse than the typical "Haskell is {only for academics, not suitable for applications that need to handle mutable state, right-wing}" that people like to throw around.
Well, you can't expect that, but what I am saying is that it is a blog, and I think a blog author should be free to write for his audience. Parent comment seems to say that you should write every blog post as if it is intended for the general public, which I disagree.
No, that's not even remotely close to what I said.
I'm saying it wouldn't hurt his case to write a small blurb about him being one of the designers of SML in the sidebar or to use language like "when we designed SML" or "the choices we took in SML," so it doesn't appear as if it's a wholly objective critique or even subject line.
I get an odd taste in my mouth seeing a core language contributor bashing other languages for being "inferior," whilst attempting to appear neutral.