Microsoft brain fuck

Been trying to get a .NET 3.5 demo app up and running for the past couple of days.

It turns-out that as far as IIS is concerned there is no .NET 3.5, just .NET 2.0 apps with a bit of extra cruft.

So, to get this working I had to create a ASP.NET 2.0 Application Pool in IIS, deploy the 3.5 app from VS2010, go back to IIS & change the app's Application Pool, AND THEN... because the template's web.config was hosed... comment-out a bunch of lines in the app's web.config.

Like I said, brain fuck.

vim for VisualStudio 2010!

http://github.com/jaredpar/VsVim/

Oh my :-0

To get bollocked by Dave Korn... my water almost came-out my nose.

Thanks for fighting the good fight Dave!

WTF?

Today I had to create a Windows Service wrapper for a Java program.

I used Apache Commons Daemon

Damn thing wouldn't work. The wrapped app was checking for command line args and kept throwing an error.

Finally I got pissed-off and added a "-fubar" command line arg. Fuck me blind if it didn't work!

If you can't be good be lucky!

R.I.P. Numb3rs

Numb3rs is cancelled.

It was good while it lasted. (Ok, not really, but I enjoyed it.)

Can't beat a show where the smart guy gets the beautiful and very smart woman.

Einstein, Dirac, Godel, Selberg, Harish-Chandra in Princeton

Even really smart guys can't decide where to go for lunch! :-)

Shhhhhhh... don't tell

I spoke ill of Java too soon.

I'm growing fond of the new iterators (but who was the moron that decide ':' was the operator?) and rather loving generics.

I'm falling in the Old Ways rapidly... time to blast out into some OCaml or F# :-)

Back to the Java ghetto

Had to write a Java program yesterday, I forgot how painful Java is compared to Ruby... even the Java 6 cruft doesn't help much... somebody please pay me to write Ruby again!

Meeting stopper

Run this during your next meeting, should move things right along :-)

Time to grow-up

This sort of thing is becoming a problem in tech, maybe it's time to grow-up.

From my point of view it's an unhealthy mix of of hyperbole ('rock-star', 'edgy', 'hero'), Adult Swim, and juvenile humor movies (someone at the Paying Day Job seriously commented that 'Van Wilder' is a 'classic' movie).

I'm not big on organized religion but this seems particularly apropos:

When I was a child, I spoke as a child, I understood as a child, I thought as a child: but when I became a man, I put away childish things.

Grow-up guys.

Daily math

Started reading Goldblatt's "Topoi" this morning waiting for someone to let me into the Paying Day Job.

Just into the Intro but it's quite good.

The book is available online (Cornell I believe), but the formatting is munged and no diagrams, so the hardcopy is nice.

The Aging Programmer

I commented on this blog entry this morning.

This sort of thing pops-up now and again with various degrees of success.

One the one hand I've felt some of the ageist pressure, on the other most of the analysis is flawed and overly emotional.

At the end of the day it boils down to what have you done lately and what can you do for me now?

Appropriate use

Over the weekend I sent a link to MagLev to the CTO of the Paying Day Job.

His reply was

Remind me to look at this later when we are focused on tuning the system...

to which I replied

Putting performance and scalability off to 'later' is not a good idea. Performance and scalability are by design, not a bolt-on.

Monday this came-up with one of the Architects who brought up the optimization canard.

When building things in the real world various materials have an appropriate use. You can use materials inappropriately, however there is usually a price to pay.

In computing we tend to ignore this for a variety of reasons. Generally I attribute this to the plasticity of the media and the fact that there's not a lot of difference between one tool and another.

However, when you pick the slowest language in the Shootout and then couple it with a single-threaded framework (we're frozen pre 2.2) you know that you're not building a system that can handle 100's of transactions per second.

You can build a F1 car on an F-150 chassis, just don't expect to be competitive.

RESTful Java

I'm sure I'm the last to know but I'm finding Jersey pretty cool.

A couple jars including the Grizzly NIO http server and I've got a no muss no fuss RESTful app server that doesn't take ANY xml to get up & running.

Annotate (which I love), code, and run.

Self (prototype objects)

Nice blog on Self.

So many languages, so little time.

UNIT (uber nerd in training)

This is granddaughter Camryn browsing my tech library... she immediately goes for Stevens' UNIX Network Programming and TCP/IP Illustrated... a solid background in the classics is central to any tech education.

She starts Category Theory next week :-)

Design methodolgy

A compliment to HDD... DDD

Saturday digest

Some cool links for today:
Making mutable state in OCaml faster
Comparison of some Ruby Mock libraries
More IRB customization
Google I/O sessions & slides
Provable software
Graph Theory (pdf)
Google Wave blog
Blackbox testing for Rack-based apps (including Rails)

Tachyon debugger

Saw this old post about odb this morning, makes me think Ruby should have something like it.

I'd call it the Tachyon debugger, which of course brain links to one of the best bad movies of all time.

The neurons are firing on all synapses this morning! :-)

Nice IRB add-on

This is nice for those of us who refuse to leave the command line.

One editor to rule them all...

Another decent intro to using gvim as your ide. This if for python but would work equally well for ruby.

Toss this in for debugging and you're off to the races.

I love vi, vi loves me :-)

Morning Math

I love the smell of Math in the morning ... it smells like VICTORY.

Sadness

A moment of silence and all flags at half staff please...

Drunkard's Walk

I put The Drunkard's Walk into my library queue this morning, it got me to thinking about my issues with the current fads (TDD, BDD, Agile, pair, ...) in software development. I'm wondering if our numerous false conceptions about the way things work aren't at the root cause of the problem.

Brains are certainly tricky devices :-)

Testing

We're using RSpec & TDD on this engagement.

TDD is no great shakes, the cart before the horse imo but to each their own.

I am increasingly worried about what specs 'prove' (in the loosest sense) about the correctness of code. And even more about what they say about code as it changes.

I've looked at generating tests (and correctness 'proof') from the AST via RubyParser but even a simple class here has at least several mixins and before/after filters. Given Bob Martin's RailsConf snark about Smalltalk's "death" ("it's too easy to make a mess" (via Ward Cunningham)) that's definitely the pot calling the kettle black.

We'll see where this goes.

Back

Wow, looking at that last post it's been awhile, lots of water under the bridge.

Spent most of last year working on SolarPowerMe (ActionScript, Ruby, Rails, WebOrb). We then took those ideas and simplified them and moved them to Facebook (all Ruby, Rails, and Facebooker)

That didn't pan out so now I'm on a Ruby/Rails contract gig in KC.

Ruby & Rails is fine but really nothing new (for an old Smalltalk guy :-)... I'm working on some ideas that interest me in OCaml... more about those here later.

Pair programming

I've come to the end of my patience with pair programming.

I can't say whether or not pairing is a good idea in general, for me personally it's a bad idea.

I could be glib and say watching someone else program is about as satisfying as watching someone else have sex... but there's more to it than that.

The communication pairing requires is antagonistic to the deep thinking that programming generally requires (for me). I suppose it's another symptom of the 'multi-tasking' attitude that gives most things (driving, conversation, programming, thinking) such short shrift.

Anyway, on to the next silver bullet! :-)

Sitescope bug?

Using $$SERVER_NAME$$ in the Server slot of a WebSphere PMI template in Sitescope generates what looks like in the log as a network lookup error. I'm assuming this is a bug and $$SERVER_NAME$$ is not being substituted when the template is deployed. Grrrrrrrrrrrrr

WebOrb & .rb naming

Rails' annoying underscorification doesn't seem to work with WebOrb. It looks to be a good idea to have the class name be the same as the filename... which is as it should be.

WebOrb & Rails 2.

To connect WebOrb to Rails 2.x make sure to disable forgery protection in appropriate config/environment/*.rb file:
config.action_controller.allow_forgery_protection = false


Problem with Gem on Windoze

Last night trying to install Rails on Windoze I ran into the dread:
ERROR: While executing gem ... (Gem::RemoteFetcher::FetchError)
getaddrinfo: no address associated with hostname.(SocketError)
This seems to be fairly common, with no solution. It turns out (after hacking remote_fetcher.rb) that HTTP_PROXY was set in the environment (the day job), unset it and everything works fine.

You gotta love a language that comes with sources, programmer heal thyself :-)