--------------enigF762B0C76BF662A72DCDC507
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

Sam Smoot wrote:
> David Vallner wrote:
>> Like integrated unit test support.
> 
> Meh? What's a couple keystrokes? Does that really help you be more
> productive?

Yes. I've switched from Eclipse and 250 dollars + VAT to IDEA precisely
because they add up. It's also not only about the less time spend
typing, but about keeping the flow of your development cycle fluid and
without mental context switches.

>> Project management.
> 
> Why would I want Gantt charts in my IDE?
> 

Not THAT meaning of project management. I meant keeping track of what
files belong to which project, or keeping track of interrelated
projects. Being able to only specify this dependency and never have to
bother with library load paths is one more thing utterly unrelated to
the code itself that you don't need to keep in your head, which is a
Good Thing.

>> Semantics-aware code templates.
> 
> You don't need an IDE for this.
> 

An IDE implemented in Emacs Lisp / vimscript still is an IDE. Though
this specific feature doesn't have much to do with the main focus of
IDEs, that's providing uniform access to a set of tools, it is true that
what you commonly call IDEs usually implement more extensively
code-aware editing features.

>> Erm. The thing is you pick an IDE, stay with it, and use one way to
>> consume those things only instead of keeping idiosyncracies of
>> command-line interfaces in your head.
> 
> Lucky you. I'm not seeing how "IDE idiosyncracies < command-line
> idiosyncracies" though. Plus I work on Linux, Macs and Windows. So a
> unified set of tools available on every platform is really nice.

Surprisingly enough, cross-platform IDEs are not unheard of in the 21st
century. It's also missing the point, the issue isn't that a given
toolset isn't available on a platform, rather that there is an impedance
mismatch between the different tools. How easily can you look up
documentation for some element you see in the debugger or irb? I
consider a copy/paste operation as something that takes too long for this.

>> Also, this is too strongly an issue of personal work patterns to be
>> preached, and I find it irresponsible to mention that without a big YMMV
>> disclaimer.
> 
> Fair enough. YMMV. It's just opinion. It's not like it's going to hurt
> you to learn the command-line tools. Frankly, any Rubyist that doesn't
> know how to use "ri", "irb", "gem_server", etc is doing themselves a
> big disservice IMO, IDE or not. Whatever floats your boat though...
> 

No, it's not going to hurt. I know how to use them, I know how to use
their equivalents in all the other programming languages I regularly
use, I just find them suboptimal. I also use ruby-doc.org instead of ri
because of the cross-referencing.

> I think the reason why we don't see much refactoring support is that
> it's just not needed. That's just my opinion tho'.
> 

No, it's just mind-bogglingly hard to implement. Refactoring means the
operation must be nondestructive with regards to semantics, and it's
very difficult for a program that isn't a Ruby interpreter to
sufficiently reason on Ruby code. And because of late method binding,
even the interpreter doesn't know ahead of time for example what a given
method call might invoke. But that's digressing.

David Vallner


--------------enigF762B0C76BF662A72DCDC507
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: OpenPGP digital signature
Content-Disposition: attachment; filename="signature.asc"

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (MingW32)

iD8DBQFFhbHDy6MhrS8astoRAudCAJkB07lBuxbEm/oFocS2bQ4vByxqsgCfTcqb
LHt9yldUg9tNgxQtskhsixwza
-----END PGP SIGNATURE-----

--------------enigF762B0C76BF662A72DCDC507--