Rhonabwy

It’s the end of WWDC, and I’ve made it back to Paul’s place, taken some aspirin, and am starting to chill for a bit. The conference has been good, the people great, and I got everything I wanted from it. I’m now also completely exhausted and ready for some quiet. And maybe a little less drinking in the evenings.

The sessions were way more packed than previous years, and the “herding” was getting a bit annoying by the end of the conference. It also ended a little earlier than I recall (1pm on Friday?) - I wouldn’t have minded a little more stretched out and smaller setup. Still, it was good information overall, and the Leopard preview will be giving me things to fiddle with and think about for months coming.

Not sure what the airport will hold for us tomorrow, that’s something I’m not looking forward to. But at least I’m going to have it better than those poor folks doing the international travel.

Tuesday night and I bailed out of the parties tonight. Gus was planning on taking an early evening, but since he’s not back yet, I’m guessing he changed his mind. Heh. Last night we were out pretty late with Buzz’s weblogger dinner (hosted by NewsGator!!!)

I installed the Leopard developer preview to get a look, and the features are indeed really interesting. It’s clearly not fully baked, but I wouldn’t quite expect it to be with a spring 2007 release date as announced. It is way usable though, and I’m really rather tempted to drive it quite a bit and see how it goes.

The sessions have been excellent, although the size of the event is clearly straining some things. We’re all being pretty heavily hurded about since there’s just so darn many of us, and the really popular sessions are quickly going standing room only. Network access has been intermittent, but mostly reliable (except where you wouldn’t expect it - like a room with 700 folks in there at once). I didn’t take the laptop with me to the sessions today, just to give it a break from being hauled all over. I did take the moleskine and a pen though I didn’t end up taking all that many notes.

There’s a pretty hefty contingent of the XCoders here, so we’ve been checking in and sharing details about sessions between us.

Well, enough for tonight - since I can’t really share any good juicy bits anyway.

I was heading to write about asynchronous circuits, and in the shower this morning I got to wondering if we even had the “right math” to talk about meaningful and useful things when it came to circuitry that is inherently asynchronous. We’ve been locked into this clockworks lock-step digital stuff for quite a while, and there really was a few other choices if you take the computing field back far enough.

So I had this professor - as I recall he was never anything higher than an associate professor - name Ed Vredenburg. I can’t find a single link on the web about him - but man, he made an impression. 20 years ago, he taught myself (and a pile of other EE students) basic circuits with Dr. Leavene (Dr. Leavene was also known as “Dr Death” - although I never perceived his as such - but that’s another story.)

Anyway, Vredenburg was this gruff, smoking, kind of nasty old fellow who’s office was a 6×9 cinderblock firetrap. I don’t recall if he died shortly after 1986, but I do recall he wasn’t around by the time I graduated in 1990. To be honest, I’m not even sure I’m spelling his name right. But at the time, he was all practical and very “EE” focused - on all those things that actually rather bore the crap out of me - power engineering, transmission, etc. But for all this time, I do recall him being somewhat put off by the lockstep digital age, and I wonder now in retrospect if he didn’t have a lot more knowledge that he never could get out to anyone else. I specifically recall some disgruntled lectures/conversations in labs where he was likening your standard transistor circuits to integrators and mathematic calculations, and how he said “any of that math you come up with from calculus, you can devise a circuit to do”.

At the time, I thought “Well, that’s just all and goody, but what the hell good use is that? I’d spend 6 hours putting together these circuits when I could spend 30 minutes on paper (and maybe a calculator) and solve the problem”. Digital seemed just fine to me. I think that master of the OpAmp (which meant more to me as a means of getting a decent radio amplifier working) may have actually been a fan of the analog computer. Although it’s taken as a sidenote of potential historical interest, I wonder if we didn’t miss something crucial back there. There’s still potential thought to that, although little has hit the news recently. An article in the EE Times from 1998 is the most recent I could quickly spot.

But back to Ed - when I was looking for a link, I went to MU and looked through this faculty pages. I was a tad shocked to see all my professors in the Emeritus list these days - with a few exceptions. There’s no link to my favorite professor - Hugh Graham. Heh - he dressed a bit like a biker, there was a rumor he’d ridden a motorcycle in to his office (and the 3rd floor of the EE building), and he was a seriously smart, no-nonsense guy that knew digital computer layout extremely thoroughly. His class on VLSI design stuck with me for years. In fact, I still have the book. Other professors I didn’t care so much about - the Charlson’s held the majority of the political power, pulling in grants for semiconuctor research into gallium-arsenide chips and I always rather though of Elaine Charlseon as rather snooty and bitchy. My advisor, Dr. Deavany is still active faculty aparently. He spent most of his time way in his head - so much so that I remember watching him run into a tree (nice black eye from that!) while walking across the quad thinking about something. The world around him just didn’t really impact him, well - until it really impacted with him.

Well, enough opinions and rumors about retired MU faculty.

I’m of the opinion that I don’t have the right language, or the right “way to think” about these asynchronous circuits, and that the cycles that can happen - the feedback loops and chaotic potential - are really somehow quite critical to use performing some of these more analoge tasks better. Image recognition, voice recognition - it’s all analog stuff. The ‘pedia has some interesting backgrounding around the analog computer, but I think the important bit isn’t the specific application of any one thing, but the “how to think and talk about it”. I don’t know - maybe it’s out there, and I just haven’t heard about it or didn’t grok it when I was exposed to it.

Update: I found a link to Ed Vredenburg. Tenuous, but there. His full name was Edward J Vredenburg - and the only link I found was in the Columbia Missourian archives - he was a pall bearer in a 1967 funeral of Roger Kyllonen. And the worst is that archive doesn’t include 1986 through 1992 - pretty much my time at MU. Maybe sometime…

Gotta love Flickr!

We went a saw the blue angels in seattle - actually lake washington - this morning. Makes for a very full day, but it was neat seeing them. Years ago we’d gone down to watch them from the I-90 bridge, and today we did it again. As I’m writing this, Flickr has some 302 photo’s already posted from the show.

I was hoping to find a pic from when the two angels were flying at something like 30 mph - with their noses 60 degress up and basically in a beautifully controlled powered stall. Didn’t spot one. But the Flickr link above has a number of good shots from today and the past few days as they’ve been performing for SeaFair.

I’ve completely a nice django development image, and when I compressed the whole kit down into a zip file, it still rolls out at 1.2Gb. I’m working with Jacob to get it hosted via BitTorrent. You know, it’s actually a little scary thinking about putting a 1.2Gb file up somewhere for folks to download - I can just imagine this terrible bandwidth bill from it… made even worse if your hosting provider is actually connected as good as they all say they are.

So in the interest of sharing, I figured I’d post the “recipe” I used.

First thing I started with: VMWare Server. (It is free, but requires registration). Then I snagged the ISO images for Ubuntu 6.06 (Dapper Drake). TO get them yourselves, head to the download page, pick a mirror, and grab an ISO (I snagged ubuntu-6.06-desktop-i386.iso.torrent).

Fire up VMWare server, make a new machine, and set the “CD” to the ISO image. From there, it’s really darned easy to get yourself a quick desktop (turns out the install ISO is actually a liveCD too, so you can get a quick feel for the desktop…).

Once you’ve got a machine made, the work starts. First thing I did was get the updates. You can do it through the UI, but I just went straight to the shell for most of this:

sudo bash

apt-get updates

apt-get dist-upgrade

apt-get install ssh subversion

To install the VMware Tools, you need a compiler and the linux headers. From the updated system (after a reboot), I ran “uname -r” to see what version I had, and then I went and installed those headers…

apt-get gcc make linux-headers-2.6.15-26-386

From there, you can choose to install the VMWare tools from a menu in VMWare server, which drops a virtual DVD onto the desktop. I copied the file VMWareTools-1.0.0-28343.tar.gz to the desktop.

cd ~/Desktop

tar xvzf VMWareTools-1.0.0-28343.tar.gz

cd vmware-tools-distrib

./vmware-install.pl

cd ~

rm -rf vmware-tools

Onto the infrastructure bits:

apt-get install apache2 sqlite3 sqlite3-doc python-pysqlite1.1 python-mysqldb mysql-server mysql-client python-docutils py-mysql python-psycopg postgresql-doc-8.1 postgresql-8.1

And finally - the Django bits themselves…

Download the tarball: Django-0.95.tar.gz

cd ~/Desktop

tar xvzf Django-0.95.tar.gz

cd Django-0.95

python setup.py install

The last bit - I did with the UI. I like Stani’s Python Editor fairly well, so I installed it too. With Ubuntu, it was pretty darn easy:

Choose Applications… -> Add/Remove…

A dialog box opens and loads up everything it knows about. I check the “Show unsupported applications” checkbox, and select the subject “Programming”. Scroll down the list until I find SPE (version 0.8.2a+repack-0.1 as of this writing), hit the checkbox, and click OK. Away it goes. And year - it takes a while.
That’s it - the recipe. Nothing’s configured, but all the components are in place.

What? Oh hell - I don’t have any.

Okay, I can make the obvious ones - like they’ll show off leopard (some prediction, it’s only been announced…) and they’ll take the “oh, we didn’t actually mean to release it” DashCode (which is kinda neat - it was on my install DVD…)

I’m looking forward to seeing what Jens and Eric will be talking about. I’m looking forward to getting VMWare on my Mac (yeah, I’m waiting for VMWare specifically…), but I don’t expect to see anything brilliant at WWDC about it. And no, I’m not betting against Gus - because then I’d probably just owe him even MORE beer or something.

I suspect Paul is overly concerned in his fears about what Virtualization will do to the platform, I’m a little more with Daniel in his opinions. Of course, what I want it for is to run a light linux system to check through my own release processes… Yeah, I’ll get a copy of Windows on there somewhere, but frankly, that’s not as important to me. Karen on the other hand… she’s really excited about it - but at this point I think she’d be just as happy with Boot Camp.

I’m curious to see how many people bring Michael-cards to the gig. That could be really funny if worked right… I have these visions of confused-michael cards being raised in tandem at a session or two…

Ah - so that’s it for my predictions. I’m mostly just looking forward to getting back to WWDC!

UPDATE: looks like we’ll definitely hear something from VMWare at WWDC.

If you’ve got some time to read and have any interest in Mac developers and their interactions with Apple, check out Rosyna’s rant at his blog.

I think it’s the single longest blog post I’ve seen in years. It’s also bizarre and definitely not a “too the point” the post, although there’s lots of points to be made. Maybe the sheer weight of the paper it takes to print it out (29 pages with 0.5″ margins from MS Word) will help attrack attention to it - or even this post.

I was recently a bit frustrated with my choices for locking down a site with Django so that nobody could see it. Well - that’s not tree, I wanted some people to be able to see it. Just not anyone…

After looking around a while, I spotted a solution that used Django Middleware to direct anyone who wasn’t logged in to a registration page. That provided the inspiration for a slightly different solution: I wanted to lock down the site to anyone (accounts or not) except for those folks who came through a beta signup process. I didn’t nessecarily want to “register” those folks with the site - and in fact, the first cut was simply “make them give me a password to get in” sort of thing.

The result was the following code:

First - the middleware:

from webcode.fuller.settings import ACCESS_COOKIE_NAME
from webcode.fuller.views import beta_auth, teaser

class BlockingMiddleware(object):
def process_request(self, request):
ident = request.session.get(ACCESS_COOKIE_NAME, None)
if ident is None:
if "sekret_access" in request.path:
return beta_auth(request)
else:
return teaser(request)
else:
return None

from webcode.fuller.settings import ACCESS_COOKIE_NAME,ACCESS_COOKIE_VALUE,ACCESS_PASSCODE

def teaser(request):
""" just returns a bare-bones template render """
return render_to_response('fuller/teaser_block.html',
{},
context_instance=DjangoContext(request))
def beta_auth(request):
"""Displays a form template giving the user the option to provide the
'secret code' to get access to the site. This could be expanded a variety
of different ways, including checking against email, etc...     """
class BetaAuthManipulator(Manipulator):
"""A lightweight manipulator used for the sole purpose of working the
teaser/block setup. """
def __init__(self,request):
Manipulator.__init__(self)
self.fields = (
formfields.EmailField(field_name="email",is_required=True,),
)
manipulator = BetaAuthManipulator(request)
errors = {}
if request.POST:
new_data = request.POST.copy()
errors = manipulator.get_validation_errors(new_data)
#manipulator.do_html2python(new_data)
if not errors:
if new_data['passcode'] == ACCESS_PASSCODE:
request.session[ACCESS_COOKIE_NAME] = ACCESS_COOKIE_VALUE
my_response = HttpResponseRedirect("/")
return my_response
return render_to_response('fuller/beta_auth.html',
{'form': formfields.FormWrapper(manipulator, request.POST, errors),},
context_instance=DjangoContext(request))