Fred’s Blog

I recently got a feral cat (Max) from a shelter in Brooklyn to be integrated into my local cat colony. Unfortunately Max came with a plastic ID collar that the shelter people had put on him during his incarceration at the pound, and obviously this collar had to come off before I let Max out into my colony (if he got it caught on a branch he could choke to death).

So the challenge was, How to remove a collar from a feral cat that is wild, terrified of humans, and cannot be touched except under sedation?

Here’s what I did (after getting some advice from Bryan Kortis at Neighborhood Cats):

First I raised Max’s cage up on blocks so that there was about a half-inch of space beneath it. This allowed the forks of a trap divider to go down through the floor of the trap, “locking” it in place.

My plan was to confine Max in as small a space as possible on the “square” end of the trap (not the tapered end), so that I would have easier access to him.

With the first trap divider in place, I gently nudged and prodded him into the confinement space using the other divider. Fortunately he was more stubborn than violent so all I needed was a bit of patience.

Max was not very pleased with this at all.

Next I slid a quarter-inch diameter wooden dowel down into the cage, through Max’s collar, and through the bottom of the cage.

This effectively “locked” him in place, making it easier to access the collar, and making it less likely to injure him in the process of cutting the collar, which came next:

Here I was able to reach a pair of scissors in between the bars of the trap and make the cut. This was why I chose the square end of the trap: depending on where Max had chosen to position his body, being at that end of the trap gave me more options for getting at the collar. If necessary, I could even have opened the trap’s front door and reached in between the bars of the divider, which are wider than the trap’s bars.

And here, the collar is cut.

Voilà.

Since I started making my own catfood back in August, I have refined my procedure somewhat.

For one thing, I’ve become more skilled at carving up a chicken carcass. With a good heavy santoku knife and sharp kitchen scissors I can skin and dismember a 4.5 lb chicken in about 20 minutes. I got considerably faster when I discovered there are pressure points that “pop” out the shoulder and hip joints, allowing easier cutting and removal.

One of the big problems I had in my early attempts was that bones would jam the machine and bone chips got jammed between the round grinding plate and the housing. Thus not only did I have to stop and clear the bone jam every few minutes, but it was very difficult to remove the plate to clear the jam on account of the jammed-in bone chips.

I discovered the solution: by trickling water through the input tube during the bone-crushing phase, the mechanism stays lubricated and there are no more jams. This also helps to keep the holes in the plate clear, making clean-up much easier.

Cutting the bones into smaller chunks also helps a lot.

I try to buy chickens not larger than 4.5 lbs, since smaller birds have smaller bones and that makes it easier on the grinder.

So here is my process:

1. Wash the chicken in running water to remove surface bacteria.

2. Skin and dismember the chicken. Discard the skin and as much fat as is convenient.

3. Remove as much muscle meat as possible from the bones (this gets easier with practice). Separate the muscle meat and bones in separate bowls.

4. Chop the bones into lengths not longer than 1 inch. This is quite satisfying if you use a sharp, heavy knife. I always keep my unused hand behind my back while doing this so that I am not tempted to chop my fingers off.

5. Measure out the water that is needed for this weight of chicken (see chart below) and set it near the grinder.

6. Install the medium plate (5mm diameter holes) in the grinder.

7. Send the bones through. After every few bones, trickle a few tablespoons of water into the intake chute. Don’t stand directly in front of the grinder, as meaty water will sometimes spit out. Don’t wear your good tuxedo when doing this.

8. When you’re finished with the bones, put the ground-up bones aside and put an empty bowl in front of the grinder output. Flush the grinder with a good portion of the remaining water (but don’t use too much– at the end of the day, you don’t want to use more water than the recipe calls for). You now have a bowl with bony, meaty water in it.

9. Open the grinder and scoop the bone chips that didn’t go through into your bone-water bowl. Set it aside; we’ll deal with it later.

10. Now install the large plate (8mm diameter holes) in the grinder. Send the muscle meat through.

Note: The idea here is that you want fairly large chunks of muscle meat for your cats to chew on. This is very good for their teeth as it is a natural way of removing tartar. I have read some recipes where people chop up their muscle meat with a knife rather than grinding it; I tried that with my cats and the chunks were too big– they didn’t acknowledge that the chunks were actually food and they ate everything except the chunks. All my cats except Jill, who was born feral, did this; Jill (who is known to have eaten her own children– more about that another day) tore into the big chunks with relish.

11. Set the ground muscle meat aside and use the remaining water to flush the grinder; use the bowl of bone-water from earlier.

12. Open the grinder and scoop anything that didn’t grind into the bone-water bowl.

13. Pour the contents of the bone-water bowl, along with the unground bones, into a blender and run it on high for a few minutes. The result will be a nice pink mush of watery goop. If any large bone chips remain, discard them.
14. Add your nutritional supplements: egg yolks, vitamins, psyllium. Add psyllium last as it is a coagulant and if it sits too long you will be left with a firm stinky mass of jell-o that is hard to work with.

15. Mix the bone-meat, muscle-meat, and supplements in a big bowl. Decant into containers, and freeze.

That’s it. It takes me about 2 hours to process 9 lbs of chicken, including clean-up.

Here are the proportions I use:

Notes: I tried using the egg whites, but my cats couldn’t stand the taste. I’ve also read that they can’t digest them.

I haven’t been able to find the supplements in free-form, so I have to waste quite a bit of time opening up the gelcaps and pouring out the powder (or, in the case of the oils, piercing them and squeezing it out).

As for the Taste Test… my guys love it. I tried giving some to Sammy, a feral kitten I’m confining until I can get him fixed next week, and he scarfed it down and now won’t touch the canned food I give him. Picky. Sheesh.

I’ve got a script that runs through the alert logs for all the Oracle databases I manage every 10 minutes or so and sends me a text message containing the error number if it finds an error. Sometimes, when the errors are of the more exotic variety, I am left scratching my head– Do I need to leap into the batmobile and race to the nearest terminal to avert the latest crisis? Or can I continue sunning myself on the roof of my Barcelona penthouse? It is always a dilemma.

I could look up the error on one of the scads of websites out there that list Oracle error messages using my Treo’s (painfully slow) Internet connection, but that is, as I believe I said, painfully slow.

Makes me wish there was a lightweight web page out there, without images or advertising, where I could look up the error and get a quick synopsis of it.

Well now, thanks to me, there is such a web page: http://phlonx.com/oerr

It contains all the currently-documented ORA, PLS, and RMAN errors. I could add more if I find I need to add them or if someone requests it.

One nice feature of this page is that it accepts alphabetic input from a Treo keyboard. For instance, to look up ORA-16005 you don’t have to toggle your keyboard to numeric input, you can just type eg00f and it gets automatically translated.

Ironically, the Oracle error messages reside in a MySQL database. Massaging the raw data into valid INSERT statements was a bit of a chore in the beginning because I was using as source material a file that comes installed with every (Unix, seems to be lacking from Windows) Oracle installation: $ORACLE_HOME/rdbms/mesg/oraus.msg. This file is rather sloppily maintained, making it difficult to parse out the different fields in a consistent way. Took me a lot of trial and error.

This file only contains the ORA- errors, so for the others I had to turn to the official Oracle documentation itself: http://download.oracle.com/docs/cd/B19306_01/server.102/b14219/toc.htm …and reformatting the raw data in those pages was a breeze using Textpad and regular expressions. This technique missed the fields that span multiple lines, but I didn’t worry about that since those are few and could be handled manually.

Back when the tainted petfood scandal broke earlier this year, I resolved to make my own catfood. (Which is to say, I would continue to eat my own regular food, but I would make catfood for my cats).

It wasn’t until I met someone who did it herself that I was inspired to take the plunge. My Northern Industrial Tools Electric Meat Grinder finally arrived, and I took it for a test drive with a whole chicken I bought.

I’m a squeamish city boy not used to handling wildlife. Never seen a whole chicken before (except at the petting zoo). I was surprised to find a little paper bag full of organs inside the chicken– quite a convenience! I had no idea chickens had evolved to the point where their vital organs grew in paper bags.

It was hard not to anthropmorphize the chicken. Peeling the skin off was a bit of a chore. It reminded me of trying to wrestle the dress off a… well never mind.

I also discovered the importance of having a good sharp knife (I don’t). After a few minutes of hacking the chicken ceased looking like a prom date and more like Alien.

I worried about the bones. This grinder is not rated for bone crushing, but my informant assured me it handles chicken bones without a problem. So it does. First I sent the bones through using the coarse filter, but found the bone fragments a little big for my comfort. So I sent it all though again using the fine filter, so that the resulting bone fragments had a consistency of fine sand.

My recipe was simple. I added water and taurine; there are other things I could (and should) add in like psyllum, once I get them.

But the proof of the pudding was in the eating… of which there has so far been little. All my indoor cats have so far turned up their noses, but the ferals outside love it.

The problem I think is that they’re addicted to dry food. I started feeding Purina One for Sensitive Systems to Squid, who has a sensitive tummy, and when everyone turned out to like it, I just fed that to everyone by default. So training them out of the dry food habit will be a chore.

Here are links to the instructions I followed:

http://www.catnutrition.org/pictorial.php

http://www.catinfo.org/makingcatfood.htm

Someone recently asked me how to rename their C:\Documents and Settings\mysername folder but they didn’t leave a correct email address so my reply bounced. So here, for the benefit of all, is my answer.

Under the Documents and Settings is a folder saved in my name instead of my sons. When I had the new hard drive installed and XP the tech used my name instead of my son. Since it is a system folder I cant change it and really don’t want to pay another bill for trying to. Any other option?

Tosca

Tosca,

As far as I know, you cannot change the name of that folder since it is linked to your user name, and that cannot be changed. (You can change the user name alias, which appears on the login screen, but that does not change the name of the underlying folder in “Documents and Settings”).

You can, however, create a new account, transfer all your settings from the old account profile to the new one, and then delete the old one.

In detail:

Let’s say you have an account named “dad” and you want to change it to “son”.

Go to Control Panel, User accounts, Create a new account.

Type in the account name (”son”), click Next, make the account type “administrator”.

Restart windows and log in using the “son” account you just created.

Right-click on the “My Computer” icon on the desktop, choose Properties, and click the Advanced tab.

Under “User Profiles”, click Settings.

Select the “dad” profile from the list, then click “Copy to”.

Click the Browse button, and browse to the location of your new user profile (it should be in C:\Documents and Settings\son

Click OK, Yes (when it asks if you want to delete the contents of the directory), then OK out of all the dialog boxes.

Restart the computer and log in as “son” again.

Finally, you can delete the “dad” account from the Control Panel, User Accounts.
Hope this saves you a trip to the tech guy!

Here are some notes from Steven Feuerstein’s Best Practice PL/SQL seminar held at St. John’s University in New York on October 16, 2006. This is not a complete blow-by-blow of the seminar, just a few points that especially piqued my interest.

Tools Discussed

• Use of the timer object for testing PL/SQL code
• PL/SQL Developer
• SQL Developer
• Code Tester
• QNXO
• Of course, TOAD and SQL Navigator.

Code-writing methodology

• First, write a functioning code stub.
• Then define your tests using the new (free but beta) Code Tester tool. DO NOT use dbms_output to test your code!
• After you have defined all your test cases, start filling up the stub with code. As each piece of functionality is added, run your tests and proceed only if the desired output for a given test is received.
• This is a revolutionary way to write code, IMO, provided the Code Tester utility can really deliver as promised. I’ll be checking it out.

Encapsulate SQL behind a standard PL/SQL API

• Since Java is fully able to work with Oracle cursors, there is absolutely no reason to permit SQL in Java code. Hide SQL within packages. That way data model changes do not break application code.
• There have been tools for generating standard database interfaces available in Java for years (newer frameworks like Ruby on Rails provide object-oriented interfaces out-of-box, with no code generation required). Now there is at least one tool available for PL/SQL: QNXO
• Downside: changes to the encapsulating packages invalidate all other code that relies on them, necessitating recompilation.

Use bulk binding (BULK COLLECT, FORALL) instead of conventional binding (cursor FOR loops)

• Downside: can exhaust PGA
• 10g optimizing compiler converts cursor FOR loops into bulk binds (unless it is disabled via the PLSQL_OPTIMIZE_MODE parameter. It is enabled by default. To enable: plsql_optimize_mode=2)

Exception Handling and Logging

• Write your exception handling FIRST, before writing code.
• The built-in logging functionality is lame. Devise your own logging tables to capture error information.
• Log information from newer features like dbms_utility.format_error_stack
• Hide exception-handling within your API. That way, as your error logging methodology becomes more robust, you don’t have to change alot of your application code.
• Don’t hard-code your custom exception numbers (-20000 to -20999) in the code; use a package declaration to alias these with meaningful constant names. Or, dispense with the numbers altogether and just use string literals! This actually works.
• Downside to using constants: everytime you add a new error constant, ALL code that relies on the package must be recompiled.

General advice

• Always use packages instead of standalone procedures/functions (variables, constants and cursors can persist throughout a session with packages).
• Become intimate with the functions provided in the standard package.
• Use the NOCOPY hint with IN OUT procedure arguments to emulate “pass by reference” and thereby speed up getting large datasets out of a procedure.
• Avoid SELECTing INTO individual variables. Use %ROWTYPE variables instead. (if the cursor’s SELECT list changes, there is potentially less code to change).
• If you must use individual variables (e.g. you are creating a “fullname” variable by concatenating FNAME || ‘ ‘ || LNAME), declare this as a SUBTYPE and hide the definition behind a package interface. Makes code changes much easier if your definition of “fullname” changes or if the field definitions for FNAME and LNAME ever change.
• Always identify END statements with the name of the block. If it’s an anonymous block, you can name it thus:

<< foo >>
BEGIN
…
END foo;

• dbms_sql.describe_cursor tells you what columns are available in a cursor.
• Declare functions as DETERMINISTIC wherever possible, so that they can be used in function-based indexes.

After two months in the wilderness, my tutorial on the Three Normal Forms is back online.

I took it down because in the original, it had some serious flaws. While not factually inaccurate, the section pertaining to NF1 was vague and misleading, and actually belonged in the NF2 section. I have corrected and expanded the NF1 discussion, and I hope I have not obscured the simplicity of the original tutorial.

Thanks especially to Richard Moignard whose helpful feedback finally inspired me to make the changes.

Saturday morning I was jogging down South Beach a little south of the Verrazanno Bridge when I came upon a stupefying sight: a huge sea turtle had crawled up on the beach and died. It was a little awe-inspiring, because the turtle’s great size implied great age. I measured it with my steps; it was 4′ 8″ from tail to beak. I stood there for a while, wondering why nobody else around seemed interested in this remarkable creature, then continued with my run.

After getting home I decided there must be someone who would be interested in such a thing. I called the Bronx Zoo but couldn’t get past the receptionist who insisted the zoo doesn’t make house calls and she said I should call the City to get it cleaned up.

That wasn’t quite what I had in mind, so next I called the New York Aquarium and was transferred to someone in “the lab”. The woman I spoke to seemed quite interested and took down whatever little information I could give her.

This afternoon Cancer-grrl and I went back to take a look. I was a little worried that I might have dreamt the whole thing. As we approached the spot on the beach the turtle was no longer where I had originally seen him, but he had come loose and was floating with the tide just a few yards offshore. He was being buffetted pretty badly by the waves and it will no doubt not be long before he starts to disintegrate. We were able to take these pictures. This was a majestic beast and it felt important to make some kind of record of its death.

I think it was a loggerhead turtle, which has been known to frequent Staten Island.

Putty is a great if you only have a handful of servers that you connect to on a regular basis. But as a consultant with a growing list of clients you can rapidly find yourself having to deal with a larger and larger number of servers, and scrolling through that little list of Putty saved sessions can become a nightmare.

Then there’s tunnelling. It’s easy to set up tunnels in Putty, so that you can connect to a gateway and then use that connection to jump to a target server in the client’s network, but connecting is always a 2-step process. You have to hunt for and double-click on the tunnel session, and THEN double-click on the connection for the server you actually want to work on. If you’re using Linux or Cygwin you could write a shell script that would streamline things, but under Windows you’re stuck with a whole mess of clicking.

And here’s another annoyance: suppose you spend hours setting up all your Putty profiles, and then you need to transfer those settings to another computer. It’s possible, but it involves dumping and importing registry keys.

Well, life is too short. Here’s a technique for bypassing these putative drawbacks of Putty. It assumes you

1. have Putty and Plink installed; and
2. are running Windows.

I assume you have a basic understanding of what ssh tunnelling is and know how to set it up using the Putty GUI. There is a ton of HOWTO’s out there on the subject so I won’t belabour the point here.

No brain surgery is involved. We just write a Windows batch file for every connection we want to make. Then we can categorize these batch files however we like in the filesystem, make shortcuts on the toolbar, the desktop, whatever. The syntax is simple, and you only have to double-click once per connection.

Let’s suppose we want to access a corporate LAN via a gateway:
gateway.foobar.tv

and behind that gateway, there are three servers (not visible to the Internet) that we regularly work on:
dev.foobar.tv
prod.foobar.tv
qa.foobar.tv

First, for this set of servers, write a batch file that defines environment variables that are common to all the connections in the group:

localenv.bat

@set GATEWAY_USER=fred
@set PROD_PORT=3001
@set DEV_PORT=3002
@set QA_PORT=3003

The @ simply prevents these statements from echoing to the screen when they are executed.

The *_PORT numbers are of your own choosing; they are used to bind ports on your own computer to the different tunnels. To get an idea of what ports are currently in use on your machine, do netstat /ar at a command prompt.

Next, write a batch file that will launch both the tunnel and the target server connection together:

prod.bat
(line numbers added for reference only)

[1] @CALL localenv.bat
[2] start “Prod” /b “C:\Program Files\PuTTY\plink.exe” -ssh -l %GATEWAY_USER% -L %PROD_PORT%:prod.foobar.tv:22 -P 22 gateway.foobar.tv
[3] @ping 127.0.0.1 -n 3 -w 1000 > nul
[4] “C:\Program Files\PuTTY\putty.exe” -ssh -l fred -P %PROD_PORT% localhost

Line 1 loads the environment variables.
Line 2 launches a Plink connection in a separate window. Why use Plink? You could have used Putty, but Plink just provides a plain vanilla DOS window that maintains the tunnel whereas Putty is a pleasing interface that you can copy/paste in and customize to suit you. Notice the “start” command, which launches Plink in its own process (kind of like appending “&” to the end of a BASH command).
Line 3 is our way of getting around the fact that the Windows batch language does not by default provide a “wait” or “sleep” command. There are numerous ways to get around this limitation… This line basically waits for 3 seconds while the Plink window sets up the tunnel. You may have to adjust this wait time depending on network conditions.
Line 4 creates your main Putty window.

I trust it is clear from this example how to create dev.bat and qa.bat as well.

Another annoyance is the server will frequently boot you off if you are idle for too long. While there is no command line switch for sending keepalives, you can do this: Set up a new session using the Putty GUI. Go to the Connection section and enter, say, “300″ at “Seconds between keepalives”. Don’t fill in anything else, just save the session, give it a name, and then add it to Line 4 of your batch file like this:

[4] “C:\Program Files\PuTTY\putty.exe” -load “default_settings” -ssh -l fred -P %PROD_PORT% localhost

This will send a packet to the server every 5 minutes and (hopefully) keep your idle session active.

So there you have it. A convenient way to manage tunnel connections with Putty that is easily transportable from one computer to another.

As I mentioned previously, Cancer-grrl and I signed up to run in the Komen Foundation Race for the Cure (you can still make donations there! Don’t be shy!) for breast cancer research, which we did this fine morning along with some 20,000 other people.

I have never run in such a mass of people. I don’t think I have run in any mass of people, in fact, since my one-and-only junior high school track meet. It is hypnotic to gaze out and watch the sea of pony tails bobbing up and down almost in unison.

We did 5km in almost exactly 30 minutes, not bad considering she is just a week out of her last Taxol drip and fresh from a Herceptin shot. I felt a little like I was running alongside a celebrity: although there were many breast cancer “survivors” (identified by pink t-shirts) in the run, I saw very few who wore the badge of current chemotherapy treatment, the bald head.

Cancer-grrl, in fact, created a small crisis when she went to pick up her race materials. Since she’s not out of chemo yet she’s not officially a “survivor” so they weren’t sure how many survivor-year ribbons they should give her. They compromised and gave her half a ribbon.

Good day for a run. I definitely felt buoyed up by the goodwill that permeated Central Park today. As we left the Park and walked back to the car parked on Broadway the feeling began to dwindle as we mingled with regular people not wearing Komen run t-shirts, but there were still enough of us on the streets and in the restaurants having brunch to maintain the sense of positivity. Later, it diffused completely into the void.