robbat2: (Default)

For a class on Online Marketing, I had to dig up and review to articles on some elements of online marketing that interested me. I've purposefully decided to post one of my reviews here, as the article was very interesting (showing how spam can be profitable)



Bibliography - The Value of a Name - Arthur Middleton Hughes

(I'd like to add an alternative title: "Why spam is profitable." On a flight of fancy, I decided to seek out articles that provided a profitability comparison between direct mailings in the physical world, and digital direct mailings (email, possibly spam). I found this article to be an excellent discussion).

Hughes (2002) presents an analysis of direct mailing as compared to emailing, for the purpose of increasing revenues, both directly and indirectly. Hughes initially presents the problems behind existing data mining (including privacy, as relative value of information). Hughes then brings in an anecdote that spurred this paper into existence, and uses it provide an initial example on the value of data mining, as well as a simple calculation of the value of some mined data.

For the bulk of the paper, Hughes dives deeply into the calculations to find the value of some aggregate data, but then does well by tying it back together in an intelligible fashion. Hughes has a brief diversion to note that the value of the data may be significantly higher when market segmentation is brought into play, but does not expand further on this. Hughes' math continues for an analysis the value behind an improved retention rate, as well as specific marketing that is enabled by email (low cost products and last minute specials).

Hughes also gives special concern to the value of specific retention and followup communication, to help keep a customer. Hughes then also ties in the value of viral marketing with email (potential for a very high response rate), as well as newsletters. Hughes also covers a somewhat dark field of emailing marketing, namely renting out email lists. This is treading into the territory of spam, and provides a quick view of why spam is so profitable. (Hughes entirely avoids the mention that renting out lists can be a shady and possibly unethical business.)

Finally, Hughes ties up the article by noting that most companies can benefit from at least half of the practices he described to gain at least $15 of value out of every email address that the company has. Additionally, Hughes points out that it seems no company has worked out the value of emailing marketing with any degree of precision, and that further email address collection will positively impact any company.

References

Hughes, A.M. (2002). The value of a name. Journal of Database Marketing; December 2002, Vol. 10 Issue 2, p159, 16p

robbat2: (Default)
Ok folks, noon has passed now, back to the real world instead of this April Fool's-induced insanity.
robbat2: (Default)
A question that many application developers have, is how to deal with entirely stupid user input.
As a remedy, I'd like to propose a new signal for unix.
We will use signal '0', as that is currently unused (see kill -l).
The signal name will be SIGPEBKAC.

If applications decide that a user is being an idiot, and not providing sensible input, they may emit a SIGPEBKAC, and terminate.

With the possibilities of this new signal, we can extend Gentoo to the absolute cutting edge of new software, while cutting back on the number of bug reports. I will be adding MySQL-5.0 and OpenLDAP-2.3 into ~arch, and filing a bug for arches to fast-track it into stable.
robbat2: (Default)
ObjectOutputStream oos = new ObjectOutputStream(...);
...
write lots of objects to the ObjectOutputStream.
...

get OutOfMemory exception :-(.

ObjectOutputStream keeps a record of every object it has seen, and thus prevents the object from being garbage-collected, even when you ARE done with it.
robbat2: (Default)

If addiction is judged by how long a dumb animal will sit pressing a lever to get a "fix" of something, to its own detriment, then I would conclude that netnews is far more addictive than cocaine.

-- Rob Stampfli

Oh, this is sooo true!
robbat2: (Default)
I've had a bunch of Intel Dual Xeon 1U servers at work for more than a year now, and it's always been bugging me that there was a little light on the front that i've seen in a variety of different states (Green/Yellow/Orange/Red, and blinking vs. solid). The servers didn't come with any hardcopy manual, and I've been too lazy to look at the CD that shipped with them.

However, lately one of them has been giving very strange problems. Tomcat on it just keels over and dies, with a signal 11 from whatever JDK is on the machine (I've tried a few, as well as using known good system snapshots and more).

We shipped it back to the supplier, and Intel replaced the motherboard, but that still hasn't solved it.
So I set out to try and dig into it myself. I've tried everything I can on the machine: switching RAM, using a different IDE controller, using a different hard drive, running on just one CPU, but still no luck. I can't switch the power supply as it's a custom connector :-(.

I finally dug out the manual, and found these servers support a technology known as IPMI. It's basically i2c/lm-sensors on some major steroids. Gentoo didn't have any support for this stuff, but SUSE Enterprise Server and RedHat Advanced Server did. So now new 5 packages later, Gentoo does support this stuff, and my servers are spewing loads of data at me - but I still haven't figure out the problem of the server that isn't working.

Just how much data you ask?
See this sample stuff from the OpenHPI folk:
http://wiki.openhpi.org/OpenHPIWiki/OpenHPIDumps/2%2e0%2e2/IPMI/hpisensors/x346

http://wiki.openhpi.org/OpenHPIWiki/OpenHPIDumps/2%2e0%2ex/ipmidirect/hpitop/BC-T


The new packages are:
sys-libs/freeipmi
sys-libs/openipmi
sys-libs/openhpi
sys-apps/ipmitool
sys-apps/ipmiutil

I still have two things to do, to make it fully useful.
freeipmi - init.d script for bmc-sensor
openhpi - init.d script for openhpid

I'm also slightly concerned that some of the binaries might need better names, as they might conflict with other things in future:
freeipmi: /usr/sbin/sensors
ipmiutil: /usr/sbin/xmlconfig /usr/sbin/sensor

In other Gentoo news, I've assembled my own upstream package for a client to readhead(2) (see bug #64724). The new client reads from a file instead of taking it's stuff from argv, so it can be much faster. I still need to see about possibly putting together a patch for /sbin/rc to do an early call to readahead for things it uses. Looking at the bootchart results from this, the largest remaining boot-time hogs are modules-update and hotplug.

Also, since ciaranm put up his Amazon wishlist, and got a book from a grateful user, I've decided to follow suit as well, and here is my wishlist.
robbat2: (Default)
With my chronic snooze-button hitting, I really need a better alarm-clock system.
I see that MIT has very recently solved it for me!


http://www.media.mit.edu.nyud.net:8090/press/clocky/index.html



When the alarm clock goes off and the snooze button is pressed, Clocky will roll off the bedside table and wheel away, bumping mindlessly into objects on the floor until it eventually finds a spot to rest. Minutes later, when the alarm sounds again, the sleeper must get up out of bed and search for Clocky. This ensures that the person is fully awake before turning it off. Small wheels that are concealed by Clocky's shag enable it to move and reposition itself, and an internal processor helps it find a new hiding spot every day.
robbat2: (Default)
I checked the next TechWeek schedule, and it seems there is actually a 3-way conflict between TechWeek events:
INTD404 Presentations	18-Apr Monday 11:30-17:20 1105  Brady,Dobson,Dulic,Leacock
ITEC424 Event 		18-Apr Monday 10:00-12:50 535 	Kyrylov
ITEC427 Exam 		18-Apr Monday 11:30-12:50 630 	Hatala

This is really incredibly, as practically all of the 4th year IT students are in all 3 of those classes.
I thought we'd moved beyond manual scheduling, but it seems SFU is still in the Dark Ages.
robbat2: (Default)
Ugh.
That CRM paper was evil and nasty.
It's 11am now, and I've been up all night working on it.
So I'm going to get some sleep now.
robbat2: (Default)

Welcome to the most deceptive TechWeek ever.

DateTimeClassWhat?
14 March 10h00 ITEC417 Weekly assignment #3 deadline
15 March 14h30-17h20 INTD404 Project Presentation
17 March 23h59 ITEC423 Computer Simulation assignment #4 deadline
18 March 12h00 MTEC314 CRM paper deadline
1623 March 10h00-11h20 ITEC417 Sequential JPEG implementation project deadline and presentation
24 March 22h00 ITEC426 EJB project

It looks like a light load, but there is actually a LOT of work hidden in some of these items, as well as the ongoing INTD404 project, that I still need to get more development done on.

Additionally:
14 March, 19h00-23h30 - Marissa's birthday party
15 March, 20h00-23h00 - Opera (Cosí fan tutte)
I might only attend one part of Marissa's party, depends how I'm doing with Techweek.

robbat2: (Default)
This is not my final intinerary yet. I've got some flexibility still, the only requirement is that I need to be in Karlsruhe between 21st June and 26th June. Marissa [livejournal.com profile] amethest won't be finding out the final details for the Birthright Isreal flights until the end of the month. However I will need to book my flight before then.

I'd like to thank everybody that has bought the phpMyAdmin book, and donated to our SourceForge project.
We are using the collected money to organize a team meeting (paying for airfares for team members), and promote phpMyAdmin at LinuxTag 2005.

Flight: Vancouver Airport TO Munich Airport
Dep. 18h05 18 June, Arr. 13h00 19 June
Cost: 398 EUR / $649 CAD
Code: DE-6071
///
Rail: Munich Airport TO Munich Hbf
departs every 20 minutes, duration = 45 minutes.
Cost: 10 EUR
///
Maybe stay a day in Munich to see Gentoo developers (does anybody have a spare bedroom or large couch?)
///
Rail: Munich Hbf TO Karlruhe Hbf
Dep. 16h43 19 June, Arr. 19h53 19 June
Cost: $100 CAD
Code: 60-EC
///
Sleep @ Hostel (http://www.jugendherberge.de/html/jugendherbergen/individual_jh.jsp?IDJH=13) or is there a Gentoo developer in the area?
///
Rail: Karlruhe Hbf TO Munich Hbf
Dep. 12h07 26 June, Arr. 15h16 26 June
Cost: $100 CAD
Code: 65-EC
///
Marissa flys in from Israel.
///
Rail: Munich Hbf TO Salzberg Hbf
Dep. 16h22 26 June, Arr. 18h02 26 June
Cost: $50 CAD
Code: 2299-IC
///
Stay with Marissa's relatives in Bad Richenhall.
///
Rail: Salzberg Hbf TO Munich Hbf
Dep. 07h18 05 July, Arr. 09h13 05 July
Cost: $50 CAD
Code: 31014-R
///
Marissa flys back to Israel. Perhaps stay a night or two in Munich. Marissa wants to see Dachau concentration camp, and this avoids needing the extremely early morning start in Salzburg.
///
Rail: Munich Hbf TO Munich Airport
Departs every 20 minutes, duration = 45 minutes.
Cost: 10 EUR
///
Flight: Munich Airport -> Vancouver Hbf
Dep. 12h00 05 July, Arr. 16h25 05 July
Cost: 358 EUR / $584 CAD
Code: DE-2070
robbat2: (Default)
Spammers are getting less and less original, I only check my Spam box once every few weeks on average, so there are some interesting patterns that emerge.
This is a set of marginal spam, sorted by the subject line. The spam rule for 'Pharm' triggered only on occurances that had that entire string. The others scored lower, and were only caught by the body.

Now look at how their obfusciation algorithm works.
   1046 N * Mar 11 Andrew Esparza  (  22) Internet P/harmacy
   1047 N * Mar 05 Young E. Hahn   (  22) Internet P\harmacy
   1048 N * Mar 02 Lisa Kirkpatric (  22) Internet Ph!armacy
   1049 N * Mar 03 Rodger Keller   (  22) Internet Ph%armacy
   1050 N * Mar 01 Jami Hatfield   (  22) Internet Ph@armacy
   1051 N * Mar 08 Cristina Eubank (  22) Internet Ph]armacy
   1052 N * Mar 10 Arnulfo Cruz    (  22) Internet Pha(rmacy
   1053 N * Mar 11 Lula R. Chang   (  22) Internet Pha*rmacy
   1054 N * Mar 09 Clark Swan      (  22) Internet Pha/rmacy
   1055 N * Mar 07 Glen Woodruff   (  22) Internet Pha\rmacy
   1056 N * Mar 02 Hillary Abraham (  22) Internet Phar/macy
   1057 N * Mar 05 Freddy K. Stubb (  22) Internet Phar\macy
   1058 N * Mar 02 Charley Meeks   (  22) Internet Pharm(acy
   1059 N * Mar 06 Annmarie Kruege (  22) Internet Pharm*acy
   1060 N * Mar 10 Hollis Brown    (  22) Internet Pharm^acy
   1061 N * Mar 02 Sandra Barnard  (  22) Internet Pharma!cy
   1062 N * Mar 07 Johnathan W. Ba (  22) Internet Pharma*cy
   1063 N * Feb 28 Fay Kay         (  22) Internet Pharma/cy
   1064 N * Mar 01 Beverley Gibbs  (  22) Internet Pharma/cy
   1065 N * Mar 05 Deandre Carson  (  22) Internet Pharma@cy
   1066 N * Mar 08 Dexter Tatum    (  22) Internet Pharma@cy
   1067 N * Mar 03 Georgette Suthe (  22) Internet Pharma]cy
   1068 N * Mar 04 Bertha Oneil    (  22) Internet Pharmac(y
   1069 N * Feb 28 Don Neely       (  22) Internet Pharmac*y
   1070 N * Mar 07 Art Sewell      (  22) Internet Pharmac/y
   1071 N * Mar 06 Brian Ellison   (  22) Internet Pharmac[y


I think if somebody could come up with an efficent algorithm to detect permutations of a string with N characters wrong, spam detection could improve a reasonable amount.
robbat2: (Default)
Hello and welcome to the real PlanetGentoo readers.

I've been crazily busy with schoolwork lately, as I've only 6 weeks left of my final semester of university.

A number of users have been bugging me to do various things for some packages (mysql-4.1*, openldap-2.2*, autofs-4*), but I'll state right now that I'm not going to be doing those on my own for the next six weeks.
I don't feel I can provide an acceptable level of support for major changes like that on my own, and still pass my courses. This does not apply to cases where the packages are supported by a reasonable size herd that can help with my temporary lack of time, or any case where I need a package update for my job.

Reading the new PlanetGentoo, I'd like to re-iterate what ferringb said about QA & the Therac-25 problem. A very good overview of this is presented in "The Science Of Debugging" (Telles, M.A. & Yuan, H. 2001), which also covers some other major incidents, and explores the reasons behind this. The only thing I don't agree with from the book is the title, as debugging has a lot more of a holistic need to it, and also requires some intuition (we have have the meta-gcc bug for all those hardware-caused failures).

On the PHP debate raging here in PlanetGentoo, I'm a strong supporter of what Stuart said. PHP is only a means to an end. If that end happens to be the average web development project, it's a very good means, as other most other tools are too complex, or not complex enough. I've been involved in some web development projects where a tool more suitable than PHP has need, and ended up using Apache Tomcat and Apache Axis together, without any troubles (other than a steeper learning curve). There is a vague possiblity I'm biased, as I'm an upstream developer with phpMyAdmin (current rank #8 on SourceForge).
I do agree that there is also some incredibly bad PHP out there. I've seen several servers (including one belonging to a Gentoo dev [who shall remain nameless here]) get hacked entirely through PHP. Only once was this via any published exploit (phpBB related), and the remainder of the occurances were spammers purposefully probing a PHP script, and then turning a box into a source of spam. The single most common problem is stupid code like include($var); where $var has not been checked in any way, and is untrusted user input - just pass it a URL of some PHP source, and watch the server run the code. This could be stopped to a limited degree if PHP got a proper taint mode (like perl), but I don't think it would really solve the source of the problem. (I know you can limit the sources of file loading, but several of my sites need to load external material, and it is quite possible to do so if designed and coded in a secure fashion.)

Seeing some developers with Amazon wishlist links, how many devs have actually got anything they wanted sent by some grateful user?
robbat2: (Default)
Marissa recieved these emails from her father, who claims they are not rants. What do you think?

[Poll #449340]

Email #1:
Email #1 )

Email #2:
Email #2 )
robbat2: (Default)

14232.03m! w00t!.
(and a maximum altitude of 259m...)

And with that note after doing my crazy load of assignments, I think it's time for bed.
robbat2: (Default)

Play it here.
Warning, this game is addictive.
robbat2: (Default)
Ok, so I haven't provided any update on my XXS1500 porting for a while.

Not working:
USB
PCMCIA

Working marginally:
Sound

Untested:
GPIO

Working:
Networking
Flash
(and everything else I think)

Latest patch
robbat2: (Default)
This is a post to test my RSS filtering.
So that PlanetGentoo readers only get my gentoo-related posts, and no other material.
robbat2: (Default)
Sometimes cheese wheel inside earns frequent flier miles, but alchemist from always bur cashier toward crank case! beyond diskette play pinochle with alchemist beyond chain saw, or behind recliner avoid contact with wedding dress behind. Unlike so many curses who have made their smelly cheese wheel to us. Where we can hardly mourn our blood clot.

May 2017

S M T W T F S
 123456
78910111213
141516171819 20
21222324252627
28293031   

Syndicate

RSS Atom

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags