Sifu Moraga

Space Shuttle ade

2011-07-08T18:24:00.000-04:00

A Blast From The Past: Shuttle Through The Decades from NPR on Vimeo.
While looking at the short NPR video over at Vimeo of the Space Shuttle's beginnings, it occurred to me - perhaps belatedly - why this last launch is such a sad event. The whole premise of sticking yourself on top of a burning candle just doesn't compare with the elegance of returning to earth in the same vehicle you took off in with a certain degree of style. A glide landing on an airstrip is so much more civilized than a splashdown in some ocean waiting to be fished out by a glorified trawler.
So, in that sense, the Space Shuttle, did live up to it's promise. It may not have been the cheap, reusable space truck that was envisioned, but it made space travel civilized.

Reversing a binary object in Erlang

2010-12-16T08:18:00.000-05:00

While parsing a file as binary data in Erlang, which is much more efficient than as a string, which is a list of integers, you need to reverse the accumulated string. For some reason there is no BIF, built-in function, in Erlang for that. Of course you could do something like:

1> list_to_binary(lists:reverse(binary_to_list(<<"abcd">>))).
<<"dcba">>

But that is wrong on all levels of wrong-ness. Instead, searching for a better solution, I found at trapexit.org ~ Erlang ~ fastest binary reverse a code snippet that looked as promising as it was puzzling. It proved to be slightly wrong, so here is the corrected version:

reverse(B) -> S = size(B)*8, <<X:S/integer-little>>=B, <<X:S/integer-big>>.

It works by taking the binary, reinterpreting it as a little-endian integer and then converting it to a big-endian integer, which reverses the bytes. I'm not sure how fast it is, but assume the endian functionality is written into Beam, so it should be as fast as it can be.

Biking in the snow

2010-12-02T17:13:00.000-05:00

A few months ago I bought a second hand Pashley-Moulton APB bicycle as a second bike for around town and for the winter. Well, the winter came with about 25 cm of snow and today I decided to bike to work. I have Schwalbe winter spike tires mounted on the 20" wheels. So I set out but had to stop a few times to free up the gears and brakes from snow and ice build-up. Despite the spikes and deeper profile, it was a struggle through snow deeper than 2 cm, so it was lucky that most of the paths were ploughed. Ice wasn't a problem with the spikes. The trip only took 10 min longer than with my recumbent on a dry day and I had a blast, but also a workout.

Installing CouchDB on a CentOS5 system

2009-11-11T11:08:00.003-05:00

I'm migrating my new Twarfing code from home to a new server at work which involves installing CouchDB. Now, CentOS has no aspirations to be cutting-edge, so it doesn't have CouchDB in it's repository. You need to jump through some hoops to get it installed. There were good instructions on Port:EightyEight, but these turned out to be a bit out of date. This is what worked for me:

First you need the EPEL repository. I was using a 64bit system, so I did:


sudo rpm -Uvh http://download.fedora.redhat.com/pub/epel/5/x86_64/epel-release-5-3.noarch.rpm

But you might need the i386 version. If all goes well, you should see something like:


Preparing...                ########################################### [100%]
 1:epel-release           ########################################### [100%]

Next I needed the basic prerequisites for CouchDB:


yum install ncurses-devel openssl-devel icu libicu-devel js js-devel curl-devel erlang libtool

Now, this is a bit different than Port:EightyEight described it: the erlang-devel package was not available anymore via EPEL, but this turned out not to be a problem. However, CentOS did ask me to reboot at this point for whatever reason.

After the reboot, it turned out that all I needed was:


sudo yum install couchdb

and CouchDB installed without needing the erlang-lang package. Also, it wasn't necessary to install from source either. The version of CouchDB I got was 0.10.0-1.el5 which is current as of writing if you really want to know.

The local.ini file will be in a different directory if installed this way:


/etc/couchdb/local.ini

I left that the way it was for now. Let's see if something else needs to be done. The Installing_on_RHEL5 wiki page seems to be the most relevant page to a CentOS installation. In it a user is created, but yum did this already and it looks just about right:


couchdb:x:101:104:Couchdb Database Server:/var/lib/couchdb:/bin/bash

And starting it certainly worked with:


sudo -i -u couchdb couchdb -b

And


sudo /etc/init.d/couchdb start

also worked. Connecting to the web interface worked once the cluster firewall was tunneled through.

Now for testing!

I've been quiet

2009-10-23T12:04:00.000-04:00

I've not been blogging much at all recently, though I've been doing quite a bit. All the twarfing stuff takes a lot of time to develope and run. I'm also busy with some hush-hush virtualization security work. But I'll try to blog a bit more in future. I've been asked to post my malware ontology, so I'll definately post when that is up.

e-mail Truck

2009-02-10T08:52:00.002-05:00

The New York Times has an opinion article titled Plugging in the Postal Service where Ruth Goldway expounds on the virtues of converting all the US postal delivery truck to electricity. It turns out that the US Postal Service has already looked into this in 2005. They had asked a Dutch company, E-Traction, that specializes in in-wheel electric motors and did a study on exactly that. Here is the text:

At the beginning of 2005 we were approached by the U.S. Postal Service to power their small delivery vans. Our calculations showed that with relatively minor modifications an e-Traction powered version could comfortably:

1. operate a U.S.P.S. delivery van,

2. with a single direct drive SM350/1AL unit with a side mounted differential,

3. using on average 3.1 kW per hour,

4. with only eight 100 Ah Lithium-Ion batteries,

5. for an entire workday,

6. with an average speed (when in motion) of 20 MPH,

7. with a maximum speed of 55 MPH,

8. at a total distance covered of 80 miles,

9. and an overall vehicle weight of 3,334 lbs*,

10. including one driver of 200 lbs,

11. and including 250 lbs of mail.

* Based on an empty vehicle weight of 3,050 lbs. Estimated weight loss in conversion: traction motor 270 lbs, transmission 175 lbs, and differential 175 lbs. Estimated weight gain in conversion: SM350/1AL 187 lbs, batteries 220 lbs, and onboard charger 22 lbs.

The cost of recharging these batteries with cheap electricity from the power grid can hardly be compared to the ever rising cost of fossil fuel. Our SM350/1AL has been designed for 600,000 miles between overhauls. The Lithium-Ion batteries will last at least 2,000 complete cycles or roughly 8 years.

Neither article answers the question of whether our power grid could handle charging up all these electric vehicles although Goldway suggests putting solar panels on the buildings to charge up these vans and other electric vehicles that happen drop in. It's also not clear how fast these vehicles could be converted or how fast batteries could be created, seeing that Toyota is also struggling to get enough for their own vehicles.
But, it is a nice idea. I like the idea of in-wheel electric motors, as they avoid most of the inefficiencies that a traditional power train has due to friction in the gears. With the motor in the wheel, the power is transferred directly onto the pavement. The trick is in the controlling electronics which I am sure is hard to get right.
It would be interesting to find out what the US Postal Service did with the full report from E-Traction. This would be a nice little job for a budding journalist.

Save.Y.Nu

2009-01-19T16:35:00.002-05:00

Seen at the recent CCC 25C3 congress, apparently one-letter secondary level domain names are in danger and save.y.nu is calling to action. Certainly, I am in favor of shorter domain names for twitter and other applications where brevity is vital.
[Pardon the pathetic quality of my phone's camera.]

More about the IBM Chicken

2008-12-12T15:28:00.003-05:00

A few days ago I posted IBM's chicken ad without comment. In retrospect I should really explain why I thought it was significant, as well as funny. Everyone, except the German government, knows that we are in a severe recession probably by a breakdown in consumer confidence rather corporate problems, and some people are talking about depression. While, I think this has been 10 to 15 years in the making, it's a waste of time playing the blame game at the moment.
The point is that every crisis needs to be looked at as an opportunity. The Great Depression and the ensuing World War II left the US with a lot of very good things. We got Social Security, the basis for Medicare, the FDIC and other programs. World War II brought us the GI bill and allowed many people to study (like my father) who probably wouldn't have done so otherwise. This gave us a well educated population (and the US was already a leader in education back then, believe it or not) that felt well secured compared with previous generations. This, in turn, led to immense prosperity. Of course, it helped that World War II didn't destroy any US infrastructure, unlike in Europe, giving the US a head start.
So, how should this crisis be used to our long term benefit? The IBM Think marketing campaign is a good starting point for thinking about this. What IBM is trying to sell us (and yes, they want to profit it from it, of course) is the idea that technology in the hands of clever folks (IBMers, presumably, although other folks will do I guess) we can streamline our supply chains and operations. This is not only good for our balance sheets, but can be good for the environment, culture and well being. We have got in the habit of importing produce from California to New York, but a friend in California mentioned recently that they don't get to consume the produce grown on the farms next door. The supermarkets get their produce shipped in from central depots (although I hope they just getting stuff from other parts of California.)
I'd like to see people cutting the waste in the system and making these cuts permanent so that we are left with a sort of virtual infrastructure superimposed on a (hopefully) improved physical infrastructure allowing produce to reach us in an energy efficient manner. I'd also like the produce to be created in a way that makes sense. The example in the Chicken video was that it takes 10 calories to produce 1 calorie of food. Obviously, this is plain stupidity and so a cost/effect balance has to be achieved.

Surprised and Confused

2008-12-09T13:22:00.007-05:00

So, some friends have expressed surprise and confusion at my recent career changes, and all that is good as surprises and confusing events are good for the creative juices. But given that the times we live in are interesting and getting even more interesting each day, it is time to reign in the confusion a bit.
Long, long time ago, after dorking around at university, a few smaller jobs and my own company for a while, I joined IBM Research in Hawthorne, New York. Specifically, I joined the IBM AntiVirus team led by Steve R. White (get him confused with any of the other Steve Whites, and you're dead). That was fun, and lasted until Symantec bought us (or something like that) in 1998. So, I thought I'd transfer to IBM Research in Zürich (actually Rüschlikon), Switzerland to work on intrusion detection stuff (for instance Exorist) under Marc Dacier (and later Andreas Wespi) and to finish my PhD on Malware and Intrusion Detection.
By the time I finished said PhD, IBM had decided they weren't that interested in IDS after all. Of course, they up and bought ISS a year later, but who said a company has to be consistent? Anyway, I got more and more involved with compliance and security models. Interesting stuff. Yawn. No actually, it wasn't that bad and it all culminated with the design of the Data Centric Security Model, in which I was instrumental. Everyone was so impressed with the idea that they did the only thing logical: canceled the funding of the project after 1.5 years. Just in case you think I am being sarcastic, I'm not. This is a tried and proven method of finding out how important an idea is: cut the funding and see how much the researchers fight to get it done anyway. So, I left, but not because I didn't believe in the DCS model. BTW, I should point out that I can only encourage people to join IBM Research if they can (and if they are actually hiring). It was initially the most humiliating experience of my life to be around all these great minds, matched only by my attempt to improve my martial arts by joining a Chinese Opera company in Xi'an. Without any knowledge of the Chinese language. Just like with the martial arts, I thought I was pretty good before I joined IBM Research. Enough said.
I left IBM Research for a variety of personal reasons. One, though not the main one, was that I was beginning to think SOA was a good idea after all, at which point I knew I'd been there too long. So, I joined John Jay College of Criminal Justice at CUNY as an associate professor. I'm now convinced that teaching actually is my vocation, but it became clear to me last Spring that I just couldn't afford it. It was also clear (at least to me) that the US was headed into a bad recession and that there would be no easy way of compensating for the subsistence-level pay that I was getting. Landing consultancy work was going to be difficult and I found out that (quite logically really) antivirus and other security companies like to do their research in-house. So that didn't work out very well.
Meanwhile, the really big reason for leaving Switzerland disappeared. Poof. So, I contacted a headhunter in Zürich and nearly landed an very interesting sounding job as research lead. Nearly. Which means not. I had already been talking to Trend Micro and decided to fall back on their offer to join their Advanced Threat Research team. So far, I am very happy with that decision. It is a very interesting and competent team under Erik Wu's leadership. There is a lot to like about Trend. They are fiscally conservative, which normally would be an irritation, but now seems like a good thing. The CEO, Eva Chen, is very technical, so you don't have to water things down to an meaningless elevator pitch. The work I am doing is bringing together a lot of thinking I've been doing recently about semantic technologies applied to security as well as massive computational architectures. They have one of the largest Hadoop/Hbase installations while I had been dabbling with Erlang/Mnesia for very similar things.
Unfortunately, the latest career move didn't result in a physical move towards Switzerland (or even Germany). The group is actually in Cupertino, California, which would be a nice place to go to as a geek, but I'm not sure if it is the right thing for my family. However, I'm not the only team member working remotely, so perhaps this will work out well. That would be nice. I've had the luck of always doing something I enjoy, but any given job change could land me with something I would not be so keen on, so I'd rather stick with what I've got, at least for now.

I'll keep you posted. Or confused. Whichever seems best at the moment.

You are a chicken

2008-12-07T09:08:00.001-05:00

A word from our sponsors (we wish):

XPost: Announcing Local New York Chapter of the Heart project

2008-11-24T16:43:00.002-05:00

For those who don't read my other blog:
http://malwareandsecurity.blogspot.com/2008/11/announcing-local-new-york-chapter-of.html

Shearwater

2008-10-08T21:29:00.002-04:00

Shearwater is my latest music obsession AND they are coming to New York on Oct 20. I might have to cough up some dough for tickets.

(At least it isn't as crazy as Final Fantasy, Owen Pallett's band/brand/whatever. )

Twitter gives you the option to decline or decline

2008-09-23T15:35:00.003-04:00

Well, that was special! Luckily a refresh fixed the problem.

capitalism and socialism

2008-09-17T18:57:00.000-04:00

In 1991, socialism died.

In 2008, capitalism died.

RubyGems in a local directory

2008-08-24T18:13:00.010-04:00

I was trying to install gem in my home directory on the server where my site is to be hosted and read up on it in the RubyGems Manual. The system is a FreeBSD7 system running on Intel iron. So far, some things worked differently than described so I though I'd put them here for posterity:

First I created a directory for the gems and the gem program to reside: ~/lib/ruby (I already had a ~/lib/perl for perly stuff.)
Then I added the line "setenv GEM_HOME $HOME/lib/ruby/gems" to my .cshrc
Dont forget to "source .cshrc"
I unpacked the rubygem package in some directory
Next, after going into that directory, did 'ruby setup.rb config --prefix=~/lib/ruby'. (Note the lack of the trailing 'gems' directory.)
So far so good. This was more or less what the manual described. However, I had to modify the setup command to: 'ruby setup.rb setup --prefix=~/lib/ruby'. In the manual, the --prefix parameter was not there.
The same thing for the actual installation: 'ruby setup.rb install --prefix=~/lib/ruby'.
Next, ruby needed to find the gem libraries, so this line, "setenv RUBYLIB $HOME/lib/ruby/lib"
I checked the setup using "gem environment" but got an error in rubygems.rb apparently because the user_paths method was getting a string instead of an array. I fixed it by changing the use_paths method to read:
```
def self.use_paths(home, paths=[])
clear_paths
set_home(home) if home
if paths
if paths.is_a? String
    set_paths(paths)
else
    set_paths(paths.join(File::PATH_SEPARATOR)) if paths
end
end
end
```
Now "gem environment" seems to work. I'm not sure what the cause of this is and I'll leave it for the moment.
I tried to install rails via "gem install rails" ... and it worked! Apparently. Note the lack of the "sudo" before that line. This is all in my local directory on the server.

I might update this as I find more bugs or quirks.

References:

Sigg's not plastic

2008-06-04T14:50:00.003-04:00

Pallet house for sale

2008-04-12T01:39:00.004-04:00

Couldn't sleep, so I checked my RSS feeds and found this selling on Craigslist:Somehow very cool, but I don't have the land to put it on. OK, not in RL, anyway. The link on the page points to the architect's site. I still prefer the idea of building a house out of recycled containers, but this might make an interesting shed.

Quick peek at Weblin Lite

2008-04-08T12:31:00.006-04:00

Christine, of Weblin, let me know that there is an in-browser version of Weblin, called Weblin Lite. After fighting with it a bit, I got it to work with Safari and met Xantherus on the Secondlife site by using the URL http://lite.weblin.com/?http://www.secondlife.com. It was pretty fun, but I was (and am still) in a rush today, so I'll give it a more thorough spin some other time. For now, here is a screenshot from the session. Xantherus hid behind another Avatar for the shot :-)

It's a bit small, but you can see the Weblin logo on the bottom left of the screenshot and the avatars on the bottom chatting.

Open Source Lego

2008-04-06T11:16:00.002-04:00

I didn't think this would happen. I sort of thought that Lego's business model revolved around selling building plans for lego models, but apparently Lego does think that they sell the bricks and the plans should be for free. You can download recent Lego model building plans from their website for free. I guess they figure that if they mix at least one special part into a kit, they will make the sell.

I really hope they succeed as I'm a big fan of Lego and I know they've been struggling. When I told people at the recent NYC metaverse meetup about the Mindstorms kit that Sifu-let 1.0 got for his birthday I could see them see all the possibilities for RL/SL mashups. (Unfortunately, the Mindstorms is not mine to play with.)

VW2008: Virtual Worlds 2008 in New York City

2008-04-06T07:07:00.005-04:00

I attended the VW2008 as a ramp turd (that's a term from Snowcrash, look it up) at the end of the last day because it was really really expensive and I couldn't get anyone to pay it for me including myself. The idea was to hang out outside the conference itself and try to meet the people I wanted to meet outside. In the end, it turned out that the expo area was open to the public so I browsed the stalls and was able to chat with some of the vendors who were there. So this is my mini conference report, even though I missed all of the real action, which happened in the lecture halls. I'm sure someone else will be blogging about the key talks - apparently the OpenSource talk was one that would go down in history as significant.

I noticed the Sheep first, though with the lounge setup I wasn't sure w

hat they were presenting except themselves (which should be enough for them anyway). What was described as Icarus Studios on the plan was sort of a meta stall with a number of smaller vendors including Vivox. Not being that much into voice in VWs I plowed on. I spent some time with the Freggers guy, trying to explain why I was less than thrilled by the lack of first person camera mode. He was adamant that his system was accessible for the general public and though sceptical, I don't think I can speak for the general public :-) What was interesting for me what that Freggers was a German company (I picked up on his accent immediately) and this was a pattern that was to repeat itself.

Right opposite Freggers was IBM and I obviously had to drop by there. I picked up a nicely done leaflet titled the "Emergence of the 3-D Internet" a title that I totally disagree with, though I have no problem with the rest of the content. (I just don't like the term 3-D internet because I think it is misleading.) IBM is very prominent in the VW space (I may hav

e had a little to do with that when I was still with IBM) but it may not be clear what IBM is actually doing. Looking past all of its involvement in OpenSim, in its own internal VWs that no one usually sees, I think what they are really doing is creating a business market for VWs. While I'm not sure if IBM will be able to capitalize on this market that they are creating, it will be doing everyone a big favor and we should be grateful for what they are doing for us all. Any large company, IBM included, moves really slow, but it is great that IBM leadership has the wisdom to allow small groups of visionaries to press far ahead of the pack despite all the risks that this involves, especially the risk that this small group will be so far ahead of the rest that they wont be able to get the rest of IBM up to speed when the market is 'there'.

ActiveWorlds was there, the Avatar Island folks, Millions of Us, Habbo, Multiverse and others were there, too. I was running out of time, so I skipped them all as I kinda know what they do. I was told that MyCosm had some interesting ideas so I headed over that way, but all the techies were tied up and I wasn't able to talk to them. So I put myself on the beta list and I'll see what happens. What I do know about them (from hearsay) is that it is an Australian company and that they seem to be concentrating on the middleware aspect of VWs and less on the UI, although they have that as well. I'm sure that makes them very interesting for companies like IBM.

I passed by Weblin, where they were bringing the feeling of presence that we cherish in the VW to the flat, 2-D world of the browser. It is somewhat similar to Flock and it would be interesting to use the two side-by-side for a while to see which one works the best for me. However, it is a different sort of social browser I should point out. I would call Weblin an active social browser and Flock a passive one. What Weblin does is to add the avatars that are currently on the same web page as yours are to the bottom of the page and allows you to chat with them. It looks quite funky, but I can see how flock might be more practically useful as finding yourself on the same page at the same time as an avatar you really want to talk to may be a hit-or-miss proposition. I like the idea of the animated avatars on the web page as it could lead to the integration of the web with VWs at some point. Unfortunately, it is Windows only so I am going to have to wait until I can find a Windows machine.

It turned out that Weblin is a German company. Freggers is too. I constantly heard German and French being spoken amongst the participants. It looks like Europe will be playing an increasingly important role in VWs, perhaps because a greater proportion of the population has high-speed Internet access and is well-off enough to afford the services.

After the conference, we headed over to the Metaverse Meetup at the Lit Lounge in the East Village. Annie Ok had organized it and we had a great time. Many of the Metarati were there and although I wasn't able to meet everyone I was hoping to, it was great to see those I did in person. (I finally caught up with Annie, Yeah!) Being bound by the Metro North train schedule, I had to leave earlier than I would have wanted, but I was flagging a bit anyway as I had had a tough week at university.

Like many people at the meetup, I wasn't able to attend the actual VW2008 conference because it was just too expensive. Although we compensated by doing our own socializing, it would have been nice if the conference was significantly cheaper. It does sound like a good conference, though and maybe next year I'll be able to scrounge up the money.

VUs in 2008

2007-12-18T18:53:00.000-05:00

With the economic indicators looking rather bleak (at least for the US), a stream of troubling news from Virtual Universes (eg., Cory leaving LindenLabs, the Sheep firing a quarter of their flock, ...) it looking a bit dreary for 2008. So, assuming the US goes into a recession (and we wont know until we're nearly out of it again), what does that mean for us?
First of all, I don't believe that the US and the rest of the world are that coupled anymore. In the minds of many, we still are and so things wont change necessarily that quickly, but at some point I think Europe and SE Asia will wake up and discover that they are perfectly able to lead on their own.
So, there is a pretty good chance that VUs will continue to grow in stature even in a US recession. Now, the problem is that some companies will be under pressure to cut back, so we'll see if that happens.
It could also happen that with more people with time on their hands, they may try a VU out and find out what it really is about.
Notwithstanding, VUs are still pretty far from primetime. They need to be more accessible in every meaning of that word. The learning curve must be flattened. The technology more scalable (whatever happened to Distributed Virtual Worlds like Outback Online?). The technology must seamlessly mesh with conventional Web 2.0 technology (and I for one don't buy the idea that VUs are Web 3.0). The technology must be accessible from any piece of technology I have on me, phone, laptop, hud in my glasses, whatever. Etcetera.
Perhaps I'll get around to make a few real predictions before the end of the year. I'll be attending the CCC 24C3 congress and that might be a source of inspiration.
CU in-world!