Ryan`s Software
Apple II software for the 21st century

When I first started figuring out how to do Toolbox programming (and trust me, I am a rank amateur), one of the things I remember Mike Westerfield (of Byte Works fame) saying is that to do Toolbox programming, you had to use records and pointers.

“What in the name of Rennyo is a record?”

A record is like a variable on steroids. In ORCA/Pascal, it’s a “type”, meaning it’s a type of variable. What’s special about a record is that it allows grouping of several other types of variables, making it like a “super variable.” The example Mike gives is a deck of playing cards; each card has a number and a suit. To see how this would be declared in Pascal, the example goes:

type
card = record
s: suit;
v: value;
end;

Each card variable is a record, and that record has two of its own variables, s for the suit and v for the value. Presumably suit would be either spades, clubs, hearts, and diamonds, and value is between 1-13 (or, since not actually clear here, among 2-10 and Jack, Queen, King, and Ace).

Let’s use another example: Twitter. My guess is that if you’re on the Internet (which you pretty much have to be to read this), you’ve heard about Twitter by now (or you just go to this site and nowhere else). In any case, I’m continuing to try and finish up my Twitter NDA, aka IItter, which is really close to release. In doing so, I learned quite a bit about data that Twitter puts out for each “Tweet” (each individual posting to Twitter).

For instance, if we fetch some Tweets from the m.twitter.com server (the “mobile” version of Twitter, although I think this is essentially the same as we’d get from the non-mobile server), we see an individual Tweet’s raw XML would look like this:

<status>
<created_at>Mon Apr 06 09:31:02 +0000 2009</created_at>
<id>1461922309</id>
<text>Ms. Unreliable just came online. Is that time for me to go to bed?</text>
<source>web</source>
<truncated>false</truncated>
<in_reply_to_status_id></in_reply_to_status_id>
<in_reply_to_user_id></in_reply_to_user_id>
<favorited>false</favorited>
<in_reply_to_screen_name></in_reply_to_screen_name>
<user>
<id>1401261</id>
<name>Ryan Suenaga</name>
<screen_name>rsuenaga</screen_name>
<location>96744</location>
<description>Hawai'i's best known Apple II Geek and social worker</description>
<profile_image_url>http://s3.amazonaws.com/twitter_production/profile_images/91133166/rsp_normal.jpg</profile_image_url>
<url>http://uncommon-cents.net/</url>
<protected>false</protected>
<followers_count>588</followers_count>
<profile_background_color>1A1B1F</profile_background_color>
<profile_text_color>666666</profile_text_color>
<profile_link_color>2FC2EF</profile_link_color>
<profile_sidebar_fill_color>252429</profile_sidebar_fill_color>
<profile_sidebar_border_color>181A1E</profile_sidebar_border_color>
<friends_count>247</friends_count>
<created_at>Sun Mar 18 06:26:37 +0000 2007</created_at>
<favourites_count>46</favourites_count>
<utc_offset>-36000</utc_offset>
<time_zone>Hawaii</time_zone>
<profile_background_image_url>http://static.twitter.com/images/themes/theme9/bg.gif</profile_background_image_url>
<profile_background_tile>false</profile_background_tile>
<statuses_count>20632</statuses_count>
<notifications>false</notifications>
<following>true</following>
</user>
</status>

Now, that’s quite a bit of code! However, we may just want to look at a few of these elements–or at least, for IItter, I’m just looking at a few of these elements. All I’m interested in is the text of the Tweet, the screen name of whomever Tweeted, and the date and time it was created. So, we’re talking these three items from the Tweet above:

<text>Ms. Unreliable just came online. Is that time for me to go to bed?</text>
<screen_name>rsuenaga</screen_name>
<created_at>Sun Mar 18 06:26:37 +0000 2007</created_at>

And yes, I know this is not 2007, but that’s what the Twitter API returned just now!

Once I have those, I have the basis of the data I wish to display. I also know that the text of the Tweet is at most 140 characters; the created_at information is a fixed length of 30 characters (a lot like a standard Internet date and time string seen with, among other things, electronic mail); and we also know that the maximum length of a Twitter screen_name is 15 characters.

So, if we want to set up a record type to account for this information, we can do so like this:

type
twitter_record = record
text : pstring;
screen_name: pstring;
created_at: pstring;
end;

var
my_twitter_record: twitter_record;

We have declared a record type called twitter_record to suit our needs of grouping the text, screen_name, and created_at information together, then declared the variable my_twitter_record of the type twitter_record for our use in this program.

Since we know that all of the data returned are strings of text less than 255 characters in length, pstrings are quite useful here. With appropriate parsing, we end up setting our variables so that they have these values:

my_twitter_record.text is ‘Ms. Unreliable just came online. Is that time for me to go to bed?‘

my_twitter_record.screen_name is ‘rsuenaga‘

and finally,

my_twitter_record.created_at is ‘Sun Mar 18 06:26:37 +0000 2007‘

So, there’s a quick bit on records. It’s going to take a bit more time to get IItter done (I was hoping while I was off but many other things also needed attending to).

This entry was posted on Monday, April 6th, 2009 at 3:11 am and is filed under Code. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.