If one searched for Oracle and Rails they're bound to turn up a bunch of results that appear to indicate that Oracle is on the Rails bandwagon. How Web2.0 of them. (I'm trying to think of some funny snide remark about a rails acquisition, but can't really....)
Apparently someone failed to tell Oracle that the IDE for Rails, TextMate only runs on MacOSX. And MacOSX only runs (legally) on Apple Hardware. And it's been what almost two years now that Apple hardware went to x86. Oracle DO YOU KNOW THAT?
Ok, the purpose of this rant is to show how 2 sided (or out of touch) Oracle is. If you happen to need to get the Oracle Instant Client for Mac OSX to work on a MacTel machine (say if you need python or ruby or perl to access the database, or even to do Rails work), you are out of luck! Sadly it doesn't indicate that the downloads are PPC ONLY! (NOTE: This is no longer the case ;) ). But a little more digging will show that Oracle isn't planning to release drivers until after 11g comes out. (I really do feel sad for the OTN guy trying to make the MacTel users feel like Oracle actually cares (or listens). I don't envy his job.) How hard can it really be to make the drivers work on x86? Had this been an open source project it would've been done years ago. (Or I could go fix it at least). Here with a proprietary vendor, I'm at their mercy, and I can't do anything about it! How un Web2.0 of them!
(The other purpose of this rant is to note that getting python(sqlalchemy) to talk to Oracle from MacTel is a no go right now)
Good+free IDE: RadRails (Aptana), Netbeans
Good+free editors: jEdit, Komodo Edit
Commercial IDE: Ruby In Steel, Komodo IDE
'teki321' - I guess I should have stated that my whole second paragraph
was my attempt at sarcasm. It goes without saying that most Rails devs (or
at least the vocal ones) use textmate/Macs. (Perhaps they're all PPC
though...)
Of course, if you're using Rails, there's little point in using anything
besides SQLite (or maybe MySQL) since, according to DHH, databases are just
"hashes in the sky" and don't really need advanced features. It'd be a bit
like putting an Itanium in a four-function pocket calculator. Also, since
the framework handling more than a single request at a time is an outright
waste of programmer productivity, really just a flat text file would
probably scale just fine (Mongrel could serialize access to it for ultimate
performance).
Cliff - Good points. I like that DHH ruffles the feathers, though I don't
necessarily agree with all his points. Competition at all levels is good
and Rails is providing competition for many. So kudos to Rails for that.
I am the aforementioned "OTN guy" involved in the referenced discussion
thread (and there are a couple others actually).
Justin Kestelyn - Thanks for your response and reassurance that this will
eventually get worked out. My co-worker with a MacTel machine is moving
forward, since we've worked around the issue.
-State on the MacOSX Instant Client download page that it is "PPC ONLY"
-Release a non-supported, preview, alpha, beta of the Instant Client. Can it really be that hard?
-Communicate better, listening is one part of communicating
Of course Oracle loves all developer communities (C/C++, Java, PHP, Rails,
and so on); how ellse could that be?
We are actively contributing to the Ruby OCI extension (this is a recent
engagement but we are increasing our headcount in this area).
We just had a Ruby on Rails hands-on lab event, as part of the Oracle OPen
World APAC in Shanghai.
Watch for Ruby on Rails session and hands-on lab, as part of the upcoming
Oracle Open World, San Francisco, in November..
'Kuassi Mensah' - Of course you love your developers, developers,
developers. We all do. But you haven't really clarified anything that
wasn't already said by your colleague Justin. You are just asking for
people to be patient. And you are right the point has little to do with
Ruby (other than Ruby is a buzzword right now). It has to do with talking
1 year and a half to release Instant Client on MacOSX. I'm glad that you
are porting but have trouble believing that it is taking a year and a half.
It's also weird that there has to be a decision to support MacOSX (on
intel), since you already supported them on PPC?! With Oracle's experience
porting/developing running Instant Client on Sparc, Intel and PPC (not to
mention Solaris, Windows, Linux, MacOSX), I'm having a hard time believing
that it is taking so long. Certification might be another story, but I
would venture that 99% of the people requesting Mactel support don't care
about that. Release a alpha/beta/developer release/gamma (whatever you
want to call it), with no certification and let the users report the bugs
if there are any. If you cared about developers you'd give them what they
want.
In case it wasn't clear from my comments in the OTN thread that Kuassi
linked to, support for a new architecture is a business decision made by
high-level management. The decision involves a large number of variables
that neither you nor I have full knowledge of.