Never been one to let the carrier drop

http://www.sheer.us/stuff/esen/EsenMovieJam-071609R.mp3.

What is it? Well, my friend Esen is working on a movie for her parents – a bunch of photos – and wanted one of my jam sessions to go with it. So this is raw, uncensored, one pass to tape, no rehersal, improv. Hence, there are a few minor trainwrecks, especially in transpositions which I haven’t totally mastered, but this is 100% sheer jam, mostly piano with a tiny bit of panflute and pad thrown in for good measure. Please let me know if you enjoy it.

http://www.geeks.com/pix/2009/900M.html.

Granted, I already have one tiny laptop I don’t use very often (it’s a P-II).. and I’m trying to not make too many frivilous purchases.. so I probably won’t be getting one, at least not in the next few months. But, I drool..

A unusually well written overview of query optimization and indexing in mysql – http://www.informit.com/articles/article.aspx?p=377652.

Has anyone seen a much more hairy, technical article on the subject of indexes in mysql?

FSM
Fail
Fun

Here’s a couple of nifty tricks from today’s scripting adventures:

1) This one is sort of obvious in retrospect – but, if you’re processing a long list of items and want to show a status indicator, this works rather well:

for($i=0;$i<10000;$i++) {
    print "$i\r" if(!($i % 100));
}

I’ve always used the !($i % 100) trick, but never the \r, which returns the carriage to the beginning of the line.

Also fun is:

$|=1;

@spinner = (“|”,”/”,”-“,”\\”);

for($i=0;$i<10000;$i++) {
        print $spinner[($i % 4)] . "\r";
        select(undef,undef,undef,0.1);
}

2) moving a byte: handy constants to memorize are 65280, 16711680, and 4278190080, which are the second eight bits, the third eight bits, and the forth eight bits, respectively. You can do ($value & 65280) >> 8, ($value & 16711680) >> 16, and ($value & 4278190080) >> 24, respectively, to get at bytes two, three, and four. The reverse operation is even easier: $b1 + ($b2 << 8) + ($b3 << 16) + ($b4 << 24);

These lines are from a configuration file that will allow the current (4.x) Cisco VPN client to connect to the 501. They took me a couple of hours to derive – hopefully sharing this with other people will help save someone some time.

The problem: You’ve got a cisco 501, and users using the Cisco VPN client. They can’t connect, and you don’t know why. If you turn on ipsec debugging with

# logging on

# debug crypto ipsec

# debug crypto isakmp

You get things like this:

——–

ISAKMP (0): Checking ISAKMP transform 1 against priority 1 policy

ISAKMP:      encryption AES-CBC

ISAKMP:      hash SHA

ISAKMP:      default group 2

ISAKMP:      extended auth pre-share (init)

ISAKMP:      life type in seconds

ISAKMP:      life duration (VPI) of  0x0 0x20 0xc4 0x9b

ISAKMP:      keylength of 256

ISAKMP (0): atts are not acceptable. Next payload is 3

ISAKMP (0): Checking ISAKMP transform 2 against priority 1 policy

ISAKMP:      encryption AES-CBC

ISAKMP:      hash MD5

ISAKMP:      default group 2

ISAKMP:      extended auth pre-share (init)

ISAKMP:      life type in seconds

ISAKMP:      life duration (VPI) of  0x0 0x20 0xc4 0x9b

ISAKMP:      keylength of 256

ISAKMP (0): atts are not acceptable. Next payload is 3

ISAKMP (0): Checking ISAKMP transform 3 against priority 1 policy

ISAKMP:      encryption AES-CBC

ISAKMP:      hash SHA

ISAKMP:      default group 2

ISAKMP:      auth pre-share

ISAKMP:      life type in seconds

ISAKMP:      life duration (VPI) of  0x0 0x20 0xc4 0x9b

ISAKMP:      keylength of 256

ISAKMP (0): atts are not acceptable. Next payload is 3

ISAKMP (0): Checking ISAKMP transform 4 against priority 1 policy

ISAKMP:      encryption AES-CBC

ISAKMP:      hash MD5

ISAKMP:      default group 2

ISAKMP:      auth pre-share

ISAKMP:      life type in seconds

ISAKMP:      life duration (VPI) of  0x0 0x20 0xc4 0x9b

ISAKMP:      keylength of 256

ISAKMP (0): Proposed key length does not match policy

ISAKMP (0): atts are not acceptable. Next payload is 3

ISAKMP (0): Checking ISAKMP transform 5 against priority 1 policy

ISAKMP:      encryption AES-CBC

ISAKMP:      hash SHA

ISAKMP:      default group 2

ISAKMP:      extended auth pre-share (init)

ISAKMP:      life type in seconds

ISAKMP:      life duration (VPI) of  0x0 0x20 0xc4 0x9b

ISAKMP:      keylength of 128

ISAKMP (0): atts are not acceptable. Next payload is 3

ISAKMP (0): Checking ISAKMP transform 6 against priority 1 policy

ISAKMP:      encryption AES-CBC

ISAKMP:      hash MD5

ISAKMP:      default group 2

ISAKMP:      extended auth pre-share (init)

ISAKMP:      life type in seconds

ISAKMP:      life duration (VPI) of  0x0 0x20 0xc4 0x9b

ISAKMP:      keylength of 128

ISAKMP (0): atts are not acceptable. Next payload is 3

ISAKMP (0): Checking ISAKMP transform 7 against priority 1 policy

ISAKMP:      encryption AES-CBC

ISAKMP:      hash SHA

ISAKMP:      default group 2

ISAKMP:      auth pre-share

ISAKMP:      life type in seconds

ISAKMP:      life duration (VPI) of  0x0 0x20 0xc4 0x9b

ISAKMP:      keylength of 128

ISAKMP (0): atts are not acceptable. Next payload is 3

ISAKMP (0): Checking ISAKMP transform 8 against priority 1 policy

ISAKMP:      encryption AES-CBC

ISAKMP:      hash MD5

ISAKMP:      default group 2

ISAKMP:      auth pre-share

ISAKMP:      life type in seconds

ISAKMP:      life duration (VPI) of  0x0 0x20 0xc4 0x9b

ISAKMP:      keylength of 128

ISAKMP (0): atts are not acceptable. Next payload is 3

ISAKMP (0): Checking ISAKMP transform 9 against priority 1 policy

ISAKMP:      encryption 3DES-CBC

ISAKMP:      hash SHA

ISAKMP:      default group 2

ISAKMP:      extended auth pre-share (init)

ISAKMP:      life type in seconds

ISAKMP:      life duration (VPI) of  0x0 0x20 0xc4 0x9b

crypto_isakmp_process_block:src:10.101.1.61, dest:10.101.2.4 spt:500 dpt:500

VPN Peer:ISAKMP: Peer Info for 10.101.1.61/500 not found – peers:0

The solution: Try this

sysopt connection permit-ipsec

crypto ipsec transform-set default esp-3des esp-md5-hmac

crypto dynamic-map dynmap 50 set transform-set default

crypto map mymap 10 ipsec-isakmp dynamic dynmap

crypto map mymap client configuration address initiate

crypto map mymap interface outside

isakmp enable outside

isakmp identity address

isakmp policy 1 authentication pre-share

isakmp policy 1 encryption 3des

isakmp policy 1 hash md5

isakmp policy 1 group 1

isakmp policy 1 lifetime 86400

isakmp policy 2 authentication pre-share

isakmp policy 2 encryption 3des

isakmp policy 2 hash md5

isakmp policy 2 group 2

isakmp policy 3 authentication pre-share

isakmp policy 3 encryption 3des

isakmp policy 3 hash md5

isakmp policy 3 group 5

isakmp policy 3 lifetime 86400

So, I had a idea that might be a little hard to implement, but would bring some honesty back to the democratic process.

I think individual donors should be allowed to specify which campaign promise led to their donations for a candidate. Then, if the candidate failed (by a impartial observer) to at least attempt to live up to the promise, they’d have to refund the donations. This might literally mean they couldn’t run the next election cycle, because they would be too far in the hole to their donors.. but it would be a way of making sure that politicians “stayed bought”.

The problem with the current situation is that a politician can promise anything they like – and them completely forget every any promises they want after they get to office. If they upset the majority this way, they might not get reelected – but if they just upset a minority – making promises to that minority to get them extremely motivated in getting the pol elected, and then ignoring the promises after their election – nothing bad will happen to the pol, even though they’ve behaved dishonorably.

(This is partially in response to this article  – I’m disappointed, but not particularly surprised. Supporting rights for LGBTs is just not a centerist thing to do, and Obama is playing to the center, probably in the hopes of getting reelected. (says the Sheer, who is clearly not a expert at politics)

Personally, I’d rather he didn’t care about the next term – that would show that he was a truly honorable politician, willing to do the right thing on every issue even though it cost him 4 years in office*. But that’s also too much to ask for, as well I know. I guess I can always hope that after he no longer has to worry about reelection – assuming he gets reelected in 2012, and the world doesn’t end – he’ll deliver on his earlier campaign promises

* = of course, then I’d also want him to be honorable enough to bow out in favor of a Dem candidate who could win

Probably I just missed them all, but I did a fair amount of searching trying to find other users reviewing the ArcGIS 9.3 unix port, trying to decide if I should run ArcGIS on unix or Windows.

Well, I’m here to say, stick with windows. The unix port uses enormously more CPU to achive the same goals. It’s built on a porting platform called MainWin, which seems to have been written to allow porting developers to avoid having to do too much work when porting a product from one OS to another, but doesn’t seem to have given much thought to, for example, efficiency. I would guess that ArcGIS on Unix uses about 5 times the CPU – and GIS is usually pretty CPU-intensive, so this is a Very Bad Thing.

Also, a lot of functionality that works well on windows, such as starting and stopping services, deleting services, etc, works poorly or not at all on unix. ( caveat: I didn’t try installing the patch, and I didn’t try 9.3.1 – this may have all been fixed).

The unix port also definately has a ‘we didn’t feel like actually porting the application’ feel to it – processes show up with names like ‘dllrunner {hex-guid}’ and ‘arcsoc.exe’. While it’s sort of amusing watching the opposite of Cygwin happen, it doesn’t give one warm and fuzzy feelings that a lot of effort was put into taking advantage of the operating system’s native strengths.

The unix port also seems to have reletively bad process isolation. When one process is off generating cache tiles, other completely unrelated services will become unusable.

I do give ESRI props that they bothered to try and make a unix port. I just wish they had done it in the traditional way of actually porting the code / writing support libraries for unix to replace any key windows-OS functionality, instead of by slapping it in a win32-replacement framework and calling it a day.

On the *good* side – migrating the services from a unix server to a windows one took very little time, and was smooth and painless.

So, I’ve been reading about assorted musical instrument technologies that were old before I was born. I can’t remember how I got started, but the best find this time was the optigan, a 70’s era sample-playback organ that worked by storing the music optically on a disk – sort of like analog film soundtrack, only laid out like a record. Truly funny. I also got interested in the Mellotron – and reading about same – and the Hammond B3. (My ultimate studio would probably have a B3 sitting right next to a Yamaha C7 baby grand. Since I can’t afford either one, I use electronic versions instead). I learned lots of fun things about the innards of the B3 at the Hammond Wiki, including that most B3s (apparently the instruments I’ve played were atypical) have a startup sequence involving holding down a start switch until the tonewheel assembly gets up to speed, and then switching over to the run switch to let the synchronous motor take over. By the way, I still think that Boston’s ‘Walk On’ album is the final word in awesomeness when it comes to the B3. And by the way, yay mechanical tone generators!

I also read about the Ondes Martenot, a rather impressive little synthesizer for it’s day – I especially like their idea of using different speaker configurations with unusual things in front of them (gongs, sympathetic resonators) to change the character of the tone. I drooled a bit over the mighty Wurlitzer theater organ (they just don’t have enough places where a amataur can try these things out!), got a good laugh out of the Wurlitzer side man. (I’ve got to meet this Pea Hix character, if only I can convince him I’m cool enough to be worth hanging out with ;-)), and just generally enjoyed a nice bout of nostalgia for electric instruments that were old before I was born.

Those of you who follow this blog regularly will remember my nostalgia-fests over The Set, a old TV repairman’s successful endeavor to restore the first-ever NTSC TV receiver to operation, and of course my ongoing fascination (some day I’m probably going to buy one) with the CED Videodisc format. For those of you who missed out on that last one, it’s a example of geekery gone wrong at it’s finist – it was a competitor to the VCR that stored video on a disk that was read by a mechanical (or quasi-mechanical, since ideally the pickup never touched the grooves) playback head.

I also, along the way, discovered 8 track heaven, which among other things lists bands that are still releasing their works on 8-track tape. As in those clunky cartridges with a continuous loop tape inside. Oddly tempting, that..

I wonder what interesting bits of nostalgia I’ll come across next? Anyway, thanks to the authors of all those sites for keeping little bits of our geek history alive.

Oh, yeah.. I remember one thing I wanted to whine about: Geocities is shutting down, AOL hometown already has. AOL hometown – what AOL did with the free web pages hosted by all the ISPs that it absorbed in the early 90s – had many thousands of fascinating pages. All gone. AOL pulled the plug with less than a month’s notice. Jason Scott from textfiles.com did a good job of describing the carnage and also suggesting a solution that may help prevent future carnage.

Organ Stop Pizza. I wonder if you can buy the right to play the thing for a few minutes.. 😉