Gmail gets IMAP: Hallelujah!

I usually try not to blog about something everyone's screaming from the rooftops, but this is genuinely exciting: Gmail gets IMAP - Download Squad.

DS was reporting on a rumor, but they note it's official now.

I moved all my email services to Gmail lately after my longtime ISP, VISI.COM, began messing up in a big way. It took a bit of tweaking to get it all working, but it's been fine ever since. I'd have moved long ago if Gmail had IMAP.

Now I'll be able to move my wife's email to Gmail -- she likes using OS X Mail.app and she works from 3 machines. Until now I had her on VISI's IMAP service.

Big news of the day.

So, can Google's long delayed file server be all that far away?

Update: Good tip on use with Mail.app
Update 10/25: This news post has more setup tip links.
Update 11/16/07: Still more configuration advice. I still haven't gotten around to doing this, mostly because my current setup works! Still, one of these days ...

The end of the Storm worm: interesting lessons

PC World has an educational article on the Storm worm, which is now fading from the scene. Emphases mine.

PC World - Storm Worm Now Just a Squall

... Brandon Enright, a network security analyst at UC San Diego, has been tracking Storm since July and said that, despite the intense publicity that the network of infected computers has received, it's actually been shrinking steadily and is presently a shadow of its former self. On Saturday, he presented his findings at the Toorcon hacker conference in San Diego.

Storm is not really a computer worm. It's a network of computers that have been infected via malicious e-mail messages, and are centrally controlled via the Overnet P-to-P protocol. Enright said he has developed software that crawls through the Storm network and he thinks that he has a pretty accurate estimate of how big Storm really is.

Some estimates have put Storm at 50 million computers, a number that would give its controllers access to more processing power than the world's most powerful supercomputer. But Enright said that the real story is significantly less terrifying. In July, for example, he said that Storm appeared to have infected about 1.5 million PCs, about 200,000 of which were accessible at any given time....

... Since July, it's been downhill for Storm. That's when antivirus vendors began stepping up their tracking of Storm variants and got a lot better at identifying and cleaning up infected computers, Enright said.

Then on September 11, Microsoft added Storm detection (Microsoft's name for Storm's components is Win32/Nuwar) into its Malicious Software Removal tool, which ships with every Windows system. Overnight, Storm infections dropped by another 20 percent.

Today, Enright said that Storm is about one-tenth of its former size. His most recent data counts 20,000 infected PCs available at any one time, out of a total network of about 160,000 computers. "The size of the network has been falling pretty rapidly and pretty consistently," he said.

Still, Storm has had a remarkably successful run. It's called Storm because it first popped up in mid-January in spam e-mails that offered late-breaking information on powerful storms that had been battering Europe. Users who clicked on the "Full Story.exe" or "Video.exe" attachments that accompanied the spam were infected by malicious software, making them part of the Storm network.

These machines were then used to send out more spam and launch attacks against other computers. The recent MP3 stock spam that was first spotted earlier this week was sent out by the Storm network, Enright said.

Storm was effective because its creators were really good at creating messages that victims would feel compelled to click, Enright said. In its first few days, it managed to infect more than 300,000 computers, making it the worst malware outbreak since 2005. Its creators have since been masters at creating timely messages for their spam and have also had success getting victims to click on fake e-greeting cards.

The Storm network itself is constantly changing, and has used a variety of technologies that have made it an interesting phenomenon to study. In addition to the peer to peer network, it has used rootkit software to disguise its presence on the PC and a server-switching technique called "fast-flux," which makes the Storm servers harder to find on the network.

It's also developed some interesting ways of keeping researchers like Enright at bay. "If you're a researcher and you hit the pages hosting the malware too much... there is an automated process that automatically launches a denial of service [attack] against you," he said. This attack, which floods the victim's computer with a deluge of Internet traffic, knocked part of the UC San Diego network offline when it first struck.

Lately Storm has been responsible for a large quantity of "pump and dump" spam, which tries to temporarily boost the price of penny stocks. But one area that does not seem to be of interest to Storm's creators is identity theft. "Believe it or not, credit card numbers aren't worth that much money," Enright said. "It's much better to make money... via pump and dump."

It's particularly interesting that credit card numbers aren't worth stealing. Is it because there's a glut of numbers on the market, or is identify theft becoming harder?

I'm also impressed that Microsoft could knock 20% of the Storm bots offline with a single update.

Overall, this is very encouraging news. I wonder how profitable the Storm Worm really was. If these pump and dump schemes really worked that speculator behavior would come to neutralize them. (Speculators detecting early versions of the email could preempt the strategy of the scheme owner.)

Hornby on the history of Palm

Tom Hornby has written the Early History of Palm. I remember GRiD, the Zoomer and GeoWorks (wonderful software OS), but I didn't know of Hawkins role in those products.

I'm most interested in the essays to come. Will Hornby identify the critical role of Outlook's data model, and the dominance of Exchange server, in killing the Palm?

OS X 10.5 and MacTel: what the Firefox bug list tells us

Firefox has been updated for 10.5. The list iof what doesn't work in 10.5, and even in 10.4, is an interesting example of how long it's going to take to finish the Intel transition

Mozilla Firefox 2.0.0.8 Release Notes

* On OS X 10.5 (Leopard), there are known problems with some media plugins as well as Add-ons that contain binary components. Also, the tabs in Preferences > Advanced will not render properly.
* The "Close Other Tabs" action on the shortcut menu of a tab can fail with an error when more than 20 tabs are open.
* Some users have reported problems viewing Macromedia Flash content on Intel Mac computers. To work around this problem, users can remove or move the PowerPC version of "Flash Player Enabler.plugin from /Library/Internet Plug-Ins.
* Java does not run on Intel Core processors under Rosetta.
* There is no Talkback on Intel-based Macs when running natively or under Rosetta. The Apple Crash report program should launch in the event of application crashes.

OS X 10.5 radically changed the graphics layer, so we should expect lots of rendering issues for some time. I suspect that Safari 3.0 will be a better choice on 10.5 that Firefox until Firefox 3.0 comes out. That should end the Rosetta dependence too.

Interesting note on Java. Client side Java is now hopping along on one leg ....

Leopard breaks AppleWorks, what about Classic

Leopard, aka OS X 10.5, breaks AppleWorks:

Macintouch: Leopard

....For anyone using AppleWorks, convert your files before you do the migration to Leopard. AppleWorks will not load in Leopard. Some files will open in iWork & Numbers, but others will even break the new programs. Large files seem to be the issue, and converting to Excel or Word formats before the upgrade would be a smart move..."

This is a bit surprising. I wonder what other OS X applications will break with 10.5.

I'm not too surprised though, Apple has always had a somewhat cavalier attitude towards breaking applications. Too bad they don't match that with support for old file formats -- including their own dam$%ed file formats.

Which reminds me -- I wonder if it breaks Classic? Classic won't run on Intel of course, but I run it on my G5 iMac with OS X 10.4.10. There's still not much in the way of children's games or educational software for OS X [1] -- and I sometimes fire up MORE 3.1 or FullWrite Professional to open old files.

Realistically, I should wait until February until I update my G5, or until iPhone 2.0 makes me update. January, after 10.5.1 and updates to Retrospect client and a bunch of other apps I rely on, will be when I update the MacBook.

Overall I'm looking forward to 10.5.1 however. There are many things on the feature list I really want (fully supported screen sharing, remote control, iChat, Apple's new version of "Outlook" called Mail.app, the signed application model, the memory map randomization, built-in PDF manipulation, better scanning support, etc, etc.) This looks like an upgrade for power users, developers, and for the support of good things to come.

[1] Really, there's not much in the way of interesting educational software or children's games on Windows either. That market basically went to Nintendo. There's more support for older software on XP than OS X though.

Clever exploit of Apple's OS transition

Apple will provide $10 upgrades to 10.5 if you've just bought a 10.4 machine. This can be translated into a clever exploit ....

Macs-imizing your Leopard upgrade:

... If the bottom line is the bottom line, the lowest retail price for Leopard is probably $109 at Amazon. However, for a few hundred dollars more, you can get a new Mac now and a copy of OS X in the mail later that you are ethically and legally bound to install on that purchased Mac. The question then becomes: which Mac?...

The Mini is a lovely machine. My mothers has run for over a year with essentially no maintenance. I check in on it every six months or so (she lives pretty far away).

I wonder if a Mini will run OS X 10.5 server? Then you could buy the mini, buy server, put server on the Mini and that $10 10.5 goes to ...

Mindjet MindManager: If it could only do acyclic graphs

Mindjet MindManager is "mind mapping software". It lets users create an outline (hierarchy) that can render as a two-dimensional layout of boxes connected by lines.

MindManager's strengths are its Microsoft Office  integration, its marketing, its attractive output and its corporate orientation. It comes in XP/Vista and OS X versions, the latter is a true OS X app but lacks some functionality. I've written about MindManager before; functionally it's similar to the much older Inspiration but it's a lot prettier.

Pretty counts.

MindManager has one glaring defect -- from a geek point of view. It can only do trees - strict hierarchies. No networks, no matrices, no directed acyclic graphs. No inheritance.

So a box (node) can't belong to to two or more branches (arcs).

This is a pain. Any reasonably complex domain representation needs a node to have multiple memberships.

I think the UI for this is not too hard. This is basically what a "Favorite" does in XP, or a "shortcut" does in OS X. The file lives in a single place in the file system hierarchy, but a reference can appear in another place.

Symantec More 3.1 did something similar with its outlines. You could have a branch appear in more than one place. Multiple inheritance in other words. [1]

MindManager could allow users to click on a box (node) and create a "favorite" that could be dragged and dropped anywhere. They don't even need to implement full references, it would be ok if clicking on a 'favorite' merely took one to the "true" object. (Symantec MORE 3.1 did the full include model.)

If some wants to displace MindManager from my desktop, all they need to do is allow me to model an acyclic graph, or even network. Trees are very 19th century.

Ok, so they have to be pretty as well.

[1] So it's not patentable guys. It's been done.