This post inspired by Why I’m switching from Twitterrific to Tweetie.

I like some parts of Tweetie. The parts I don’t like far outweigh the parts I do. I don’t like that, while you can indeed use selecting a tweet to dim it (akin to marking it as read in Twitterrific), it automatically selects the top tweet when updating, marking unread tweets as read. I don’t like that the write new tweet box HAS to be in a separate window. I don’t like that you can’t hit the right arrow button and open the link in the tweet you’re viewing. I don’t like that if you mark a tweet as read, and it’s a mention, it’s not also marked as read in the mentions window. A tweet that’s read should be read universally.

I like that there’s multiple account access, even though that’s horribly broken. I like the URL shortener, even though it’s sometimes a URL lengthener.

Most of all, though, I don’t like that it has a menubar item you can’t turn on or off, and likewise cannot be reduced to ONLY a menubar item like Twitterrific can be.

I’m probably going back to Twitterrific soon, but I might keep using Tweetie for my secondary accounts.

For this demonstration, we’ll use @jX, @shadytrees, and @Tom. Because that’s what happened.

Tom and jX are reciprocal. Tom follows shadytrees and jX follows shadytrees. shadytrees does not follow tom nor does he follow jx, so regardless of his @reply settings, he would not see a tweet addressed in this format:

@Tom @shadytrees text

However, because Tom follows shadytrees and sees replies to people he’s following from people he’s following, both shadytrees and Tom will see a tweet addressed in this format:

@shadytrees @Tom text

I seek a smarter twitter client that allows multiperson replies (well, two is simple here, but doing it for more people that don’t have a follow network is technically impossible in a single tweet), and figures out when a person such as jX wants to reply to both shadytrees and Tom to address shadytrees first so that Tom will also see it.

So, this is the third Gawker post in a row, but it’s just so frustrating when they’re laying people off and shuttering websites when they pass up BLATANT opportunities for revenue like this:

Notice there is no referral code there. For every click though that leads to a purchase of anything in that session, Gawker Media will receive ZERO PERCENT. They do this all the time, either their sites copy a link from someone else and that person gets a bunch of referral purchases, or like this, there’s no referral code at all. They could be making anywhere from 4 to 8.5 percent of the purchase price of books, magazines, TV and movies that they link to. That’s potentially thousands of dollars, maybe tens of thousands for sites like IO9 and Kotaku. But no, no referral codes, and that’s a lot of missed money. This is why Gawker is failing, huge opportunities left lying on the ground.

Gawker, a company that’s had to fire about half their staff due to the ‘economic downturn’, has given Cheap College Gamers quite a hefty sum of money! Owen Good posted sunday coupons deals today, and linked to an xbox 360 with the Amazon referrer tag “mmbevigaotst-20″. Do a little googling and you’ll find a page on Cheap College Gamers (page linked to above) filled with links to amazon using that referrer tag.

Kotaku, a Gawker subsidiary, could be making quite a sum from Amazon commissions. Instead, some guy is making $16 (or more) per person that buys an Xbox via his link, and also cash from OTHER things people buy as long as it’s during the same visit to Amazon (which, usually, is quite a surprising amount).

I’m sure Goldenchild is very happy for the free money that you gave him, Owen Good. Nick Denton, you are a failure. (Valleywag won’t pick this failure up, for Owen writes there too)

Several of these link to fake Wii Media Download like sites which advertise Games for your Wii

You do know you can block those sites in AdSense, right? Please, do so. Also, if you wanted to make money all the amazon links in this article would have referrer tags in it. Think about it, your authors actually EARNING YOU MONEY. I know, you don’t understand things like “profit”, but you should try it sometime.

Whew! What a long title! Yep, I’m doing it to help Google out, sorry.

Long ago, I was trying to do some crazy shit with Automator and it involved Variables. I might have been doing some renaming stuff, don’t recall exactly, but I found this helpful macworld article titled THE SCRIPT IS IN THE FOLLOWING LINKED ARTICLE, COMMENTER “Gi” Automator Power: Variables and looping. It isn’t as easy to find as it should be, because its title isn’t as informative. The thing is, it does exactly what Textcast does, for free.

What does Textcast do? It apparently can turn Text, E-mail, and RSS feeds into personal podcasts (I don’t know if it has a built in webserver for RSS, which would be neat). The thing is, the automator script in the article above can get an RSS feed, and turn the text of the articles into mp3s, which are then inserted into an iTunes Playlist… exactly what Textcast does, but for $25 less. (By the way, you can have the say command piped into an mp3 file, like say -v alex -o audiofile.mp3 -f textfile.txt and it will have alex output the contents of textfile.txt to audiofile.mp3.)

Save yourself $25 then spend it by using our Amazon searchbox on the right to get something you REALLY want. For example, iLife 09.

There’s a neat thing called “Time Capsule” which Apple started pushing recently… Problem is, if you’ve already got a backup disk, you probably would like to keep your backups in order without starting anew. After all, what if you DID delete something you’d like to get back someday already? As far as I know (and I did a lot of research, but several months ago) there’s no way to port a Time Machine Backup Disk to an Airdisk backup (they use sparseimages or something).

An interesting note is, if you connect an external drive (like this My Book, similar to the one I use) to an AirPort Extreme (I have the 100mbit one, sadly), start the backup, then cancel it and connect the disk to your Mac (like my MacBook), it will start to back up to that sparseimage.

The annoying thing is, I would really like to just convert my time machine backup into one of those sparseimages, so I could turn it into an AirPort Disk and back up wirelessly. I think someone could make a few bucks selling (or asking for donations for) an app that did just that.

I called it on who would be the fifth Cylon last year. I CALLED IT!

[Edit date=2009-01-25] Please note, I mention some silly unnecessary steps in the post ahead, when I could easily set tighturl’s autoincrement to a number higher than the highest lilurl, and add some lines to tighturl to look at a different table for older (lilurl) entries. It is totally not necessary to copy over non-leading-zero ids to the tighturl table, but if this is not done, those URLs don’t get the additional data tighturl provides like hit count (unless, of course, you patch tighturl further than an ‘if less than [tighturl_floor], jump to lilurl code, else use tighturl code’, which I personally don’t think is wise, but I’m very inexperienced). Thus, while it is indeed extra work to convert valid base 36 numbers into ints in a tighturl-style table, I believe the added benefits such as hit counts and last-hit are worth it. Worth is, however, as always, determined by the user. [/Edit]

Yesterday I was bored and wanted to help Evan move ur1.ca, a url shortening service aimed primarily at identi.ca users, to TightURL. This is apparently planned, and will happen “eventually”. I loaded up the database (after messing with it quite a bit, PHPMySQL doesn’t import TSV files and saving it as a CSV was stupid as several lines include commas. In the end, I had to do a series of steps:

1. Extract the database
2. Rename it to be a .tsv and open it in Excel
3. Ctrl+F for commas and replace them with the text COMMATIME
4. Ctrl+F for semicolons and replace them with escaped semicolons (\;)
5. Save to a csv on the desktop
6. Open the csv in TextEdit
7. Ctrl+F for commas and replace them with semicolons
8. Ctrl+F for COMMATIME and replace them with commas
9. Import the CSV file (don’t use the LOAD DATA thing)

I noticed pretty quickly some oddities, like entries for 0, 00, 000, 0000, etc. It didn’t click right away that these were not base36 numbers (they’re all 0!) and were in fact just alphanumeric ids. For example, this post is http://ur1.ca/0y5s. That explained the VARCHAR(255) column (what the fuck). I ended up adding the other columns tighturl used, modifying existing ones to be proper, etc, thinking I could end up converting it pretty easily. I was quite wrong. An overhaul (and perhaps a “plugin” for tighturl, despite it’s lack of a plugin architecture) would be required.

What must be done to convert a lilurl site to a tighturl site:

1. Add a second table for entries that have a leading 0. I believe all other entries are valid numbers. This could be a big table.
2. Write some code for TightURL (and maybe get it patched in, the author seems relatively active, and he has an e-mail address) to take shorturl ids that start with a 0 and look in the second ‘compatibility’ table for these URLs created before moving to a sane system.
3. Get something that casts a fix on the column. My initial thought was to create a second column, nid (numerical id), and slowly convert each id to the nid table as an INT. I’m told this is unnecessary by people whom I believe do not know the full extent of the problem.

Of course, this could also be overcome by setting up a subdomain for LilURL and disabling the ability to create new URLs, and redirecting links that have leading 0’s there.

Evan excuses LilURL as someone’s first project, but I’m really surprised someone could think just using as-of-yet-unused alphanumeric ids was a good idea even during their first project.

I hope to add more to this post (for those who are in fact interested in doing the conversion) soon.

I’ve been reading this incredibly interesting interview with an adware author, and here’s this line that I just think is so great:

If you think about it, when I use a credit card, the security model is the same as that of handing you my wallet and saying, “Take out whatever money you think you want, and then give it back.”

Bingo. How about another model, where a credit card number functions more like an OpenID, and maybe OAuth (don’t know which would be better here), and you’re sent to your bank’s site via the credit card vendor’s redirector (preferably in a way where, like OpenID, you’re supposed to type in the bank’s site in your address bar so that you don’t get phished). Then, you are told when you log in that such and such a vendor is wanting to debit your card for this amount (once/on a recurring basis), and you approve that and are sent back to the site, kinda like with Google Checkout.

Maybe the next step isn’t OID/OA, but banks running their own PayPal like services (which will not happen for quite some time, if ever, don’t delude yourselves), and also hell freezing over.