Yariv,
Now that I realize that this is an undocumented/unsupported feature I
am a little worried about rewriting ErlyDB to take advantage of it. It
feels so right and is much more aesthetically pleasing. Would it be a
good idea to ask the core Erlang team if there is a reason they
implemented passing the module as the last argument and if it would be
possible to change Erlang's behavior to make it behave more like I
think most people would expect it to behave?
Dan
On Jun 8, 10:39 pm, "Yariv Sadan" <[EMAIL PROTECTED]> wrote:
> I must say I really like parameterized modules but I think the
> convention for passing in the "object" as the last parameter for its
> "methods" is strange. In ErlyDB as in other APIs the last parameters
> of a function are often optional for overriding default behavior. It's
> unintuitive for me to pass the most important/consistent parameter
> last when I don't use parameterized module style invocation. I'm not
> saying we shouldn't use this convention in ErlyDB because I do like
> being able to write e.g. 'Person:name("dan")' instead of
> 'person:name(Person, "dan")' but I wish this language feature were
> done differently...
>
> Yariv
>
> On Sun, Jun 8, 2008 at 12:38 PM, Mikael Karlsson <[EMAIL PROTECTED]> wrote:
> > If you proceed with on the parameterized module track you should be aware
> > that it is a not "officially supported feature" of OTP and I believe it is
> > in OTP since the R12B release and not before. Your observation concerning
> > tagging the first atom as the module could be extended to that the same
> > thing is done for records - their names are the first atom of the tuple - so
> > you can have a dual view of an "object" either as a record or as a
> > parameterized module. One thing to consider could for instance be the
> > yaws_headers.erl module in erlyweb.
> > /Mikael
>
> > 2008/6/7 Dan Bravender <[EMAIL PROTECTED]>:
>
> >> I just noticed this the other day while playing around with erlydb and
> >> I was really confused when it worked. Glad that someone figured out
> >> what's going on. It seems like quite a coincidence that the first atom
> >> in the tuple was the module name which is what the Erlang idiom for OO
> >> would require. Then again, tagging the module/type at the beginning of
> >> the tuple feels like the right thing to do (it has to go somewhere,
> >> right?) and it makes a ton of sense that this idiom exists. Really
> >> cool.
>
> >> I was sure that you had set up erlydb to work this way by design ;-).
>
> >> In the end it would lead to much cleaner code, so I'm all for
> >> reordering the parameters in the other functions.
>
> >> Dan
>
> >> PS
>
> >> Yariv,
> >> Let me know if you don't have time. I can tackle that and it would
> >> help me get up to speed on erlydb.
>
> >> On May 26, 7:34 pm, "Mikael Karlsson" <[EMAIL PROTECTED]> wrote:
> >> > 2008/5/26, Yariv Sadan <[EMAIL PROTECTED]>:
>
> >> > > Hi,
>
> >> > > I recently discovered in Erlang the capability to do OO-style method
> >> > > dispatch. It's pretty neat. I had no idea it worked but it does.
>
> >> > .....
>
> >> > Hi,
>
> >> > for those interested; there are som papers on this topic, called
> >> > parameterized modules:
> >> > Inheritance in
> >> > erlang:http://www.erlang.se/euc/07/papers/1700Carlsson.pdf
> >> > Parameterized modules in
> >> > Erlang:http://www.erlang.se/workshop/2003/paper/p29-carlsson.pdf
>
> >> > Mikael
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"erlyweb" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/erlyweb?hl=en
-~----------~----~----~----~------~----~------~--~---
