Friday, October 31, 2008

FFMPEG: All Devices

Finally found what I think VisualHub uses for its "all devices" preset:

ffmpeg -i in.avi -s 640x368 -aspect 640:368 -r pal -vcodec h264 -g 150 -qmin 25 -b 1000k -level 30 -loop 1 -sc_threshold 40 -partp4x4 1 -rc_eq 'blurCplx^(1-qComp)' -refs 2 -qmax 51 -maxrate 1450k -keyint_min 40 -async 50 -acodec aac -ar 48000 -ac 2 -ab 128k out.mp4

This preset is really good for SD content. Looks fine when I play it on my 20" iMac and looks fantastic on the iPhone too. Very little, if any, transcoding artifacts. I think for HD I will include some logic to determine if the file is indeed HD and, in that case, offer an option to encode to 720P and create an all devices version.

2008 Electoral Predictions

It's probably a little early to do these but I'm not sure I'll have time before the election. I started doing these back in 2000 and so far I've been pretty good at it. In 2000 I missed 3 states. (TN, FL, NM) In 2004 I missed only one state (OH) This year with all the battlegrounds in play I couldn't really come up with one map so I made 4. I guess that's kind of cheating so officially I will go with #2, Most Likely, as my official prediction. I would love to be wrong and see #1, Best Case of course,

Best Case Scenario: Landslide - BO 379 - JM 159

Most Likely: BO 318 - JM 220

Worst Case: BO 278 - JM 260

Racist Scum bag model: JM 281 - BO 257

Automated h264/AAC encoding on OSX

I think I've finally figured out the ffmpeg piece of my automated video encoder project. I cheated and used the Google to find some ffmpeg examples. I created a bash script front end to the command and attached it to a cron job. That works for now but it has some pretty serious problems. Mostly it can't que so if it monitors my downloads folder it just loads up a ton of processes trying to encode all the files it finds. I need to figure out a simple que system so you can set a global limit of encodes. I also want to setup an option to encode only between certain hours.

This could all be done via conf files but I'm thinking if I want an actual GUI for this app I'd rather do it on OSX. I made a quick mockup in Interface Builder for it:

I'll need to brush up on my Objective C before I can start coding anything. I know the basics but I haven't played around with it in probably 2 years. Specifically I need to know how to run external shell processes. The rest is just basic stuff, conditional checking, reading/writing preferences, etc. The other issue is how to include ffmpeg. I don't want the end user to have to install and build it themselves. VisualHub just included ffmpeg inside the app. I guess I'll do that but I'm not sure how to set that type of thing up in XCode.

Future plans might include some very basic distributed processing. Probably just something like computer 1 encodes file A, computer 2 encodes file B, etc. I think that should be pretty easy to setup with XGrid. I'm also interested in automating some meta data stuff. No clue how that would work. I would at least want to set the right type of file for iTunes (Movie/TV) and automatically insert art if a folder.jpg or cover.jpg exists.

Thursday, October 30, 2008

OSX Terminal + Wallsaver

I spend too much time in Terminal... anything to make it more exciting is like robot cocaine to me.


I'm thinking since I am going to be stuck in a hotel room most of next week I might checkout the HellaNZB source code and try to close a few active tickets for .14 I'm not much of a coder but Python is pretty damn easy. Specifically I want 7zip support and multiple growl clients. Someone already did most of the work on growl -- it just needs a couple things fixed apparently.

I also want to see how feasible it is to automatically connect hellanzb with ffmpeg. There is already post-processing support for audio so I can pretty much use the same idea. I just need to account for:

1) Known video file extensions
2) Look inside the video cointainer to find out what type of video it is
3) parse hellanzb.conf for encoder settings
4) start encoding process & send growl notification
5) check to see if the original file should be deleted
6) send final growl notification

Windows 7 UI changes

It's nice to see Microsoft has finally caught up with the NeXT GUI circa 1986. In all fairness it took Apple another 14-ish years to really get the Dock right with OSX. Still, 22 years to finally accept the Dock is a bit on the extreme side. Better late than never I guess. I've always really hated the Chicago style taskbar. As it got more and more cluttered over the years it became even less useful. This is one of those cases where Microsoft is clearly copying but you can't hold it against them. Windows users shouldn't be stuck with a outdated broken UI just because NeXT/Apple were the first company to make it mainstream. Clearly NeXT lifted elements of the Dock design from other places too. I'm just glad I can get a consistent UI experience moving from all the major platforms now.

The most interesting thing I saw out of PDC was the concept of Libraries. It sounds a lot like what I was advocating Apple do in future OSX releases. The whole file/folder/tree concept has to go. It's becoming impossible to really manage a file system on such a granular level. I don't think people are quite ready for a 100% meta-data interface but that's going to be the next step.

Overall I think they're heading in the right direction with the Windows 7 UI. The big question is if they can break all the bad habits of Windows developers and really enforce a new GUI design policy that is consistent. Right now with Vista you can see the whole range of Windows UI elements from Windows 3.11 (try to install a font via control panel if you don't believe me) all the way up to the newest .NET/Windows Forums style -- plus the various proprietary controls mixed in. It's a complete mess. At some point Microsoft just has to punish developers who don't try to buy into the consistency of a unified GUI or things will never get better.

UUID Mounting

If you've ever dealt with a significant number of hard drives in a Linux environment inevitably you reach the point where you need to move to a new motherboard or disk controller which will most likely result in something like this:

/dev/sdb becomes /dev/sde
/dev/sdc becomes /dev/sdf

If you use static device IDs in fstab this is going to reverse your drive mounts. In a server environment it could be more than an inconvenience if say your /private share is suddenly mounted on /public

There's pretty much no good reason to mount by a static device ID anymore. The best solution is UUID. Disk labels (e2label for example) would also be a possible solution however I've had mixed results with them personally so I stick with UUID now.

First, get the UUID of the volume: (for example) (also try blkid to get them all at once)
# vol_id --uuid /dev/sda

You'll get something like this: 5a4dea83-53a6-4189-8c6d-240ee7402d2d

In /etc/fstab mount it by UUID:
UUID=5a4dea83-53a6-4189-8c6d-240ee7402d2d /mnt/mountpoint ext3 defaults 0 0

# mount -a

That's about it. Not much to it. The hardest part is just breaking the habit of doing it the wrong way. It's easier up front but harder later on.

Wednesday, October 29, 2008

What would an Apple Home Server look like?

No one has really be able to bring the home server concept to the masses yet. Microsoft's product is probably the best offering so far but it's completely useless if you don't run Windows (of course, what did you expect?) I've always wanted to see Apple start connecting the dots between the cloud and multiple computers. It's a challenge that is primarily user interface based. The technology is easy -- making it understandable is the real problem.

If you haven't used OSX Server before it's basically OSX with some added functionality but with massively improved management utilities. The GUI setup tools are fantastic -- there's nothing else like it. It even includes a Dashboard widget for monitoring your OSX Server. It has the same attention to detail you'd expect from Apple. The problem is... very few people use it. It's a product targeted towards business but business runs Windows and, increasingly, Linux. I can't imagine Apple sells many OSX Server boxes. There just isn't a market for it.

Since Apple is primarily a consumer oriented company it makes a lot of sense to do a home server. More Macs being sold mean more multi-Mac households. Unless you want to manually move files there's really no good solution to keep all your data synced and backed up. It's even worse when you put an iPod or iPhone into the mix. Apple needs a solution for this problem.

Here's what I'd want from an Apple Home Server:
  • Centralized iTunes & iPhoto libraries.
  • Centralized mailboxes in Mail.
  • Multi-user support
  • TimeMachine server
  • Full Networkable home folder support as an option. (most people would be better off with syncing probably)
  • OSX network install. (excellent for MacBook Airs, forth coming Apple netbook, tablet, etc)
  • Drobo-ish setup -- easy to add, redundant, storage.
  • 802.11N + GigE + integrated router.
  • Integrate with MobileMe -- it acts as a sort of localized version of MM. It would be accessible via the internet as an iDisk/WebDAV share. MM could act as a cache -- your most recently & commonly accessed files, and as a fall-back if your home server was offline.
  • Bundle free MobileMe subscription with it.
  • Different SKUs. 2 bays, 4 bays, 8 bays, hard drives, no hard drives.
  • New feature in iTunes to connect to remote home server -- so you could be using your work computer for example and quickly sync to your home computer and grab a couple albums you forgot to sync earlier.
  • AFP, SMB, FTP, NFS (it's all built into OSX already so why not?)
  • Full web UI so you can use the device without installing extra software.
  • Reasonably priced cloud storage so the average person could backup everything to the locally and to the cloud.
  • ScreenSharing -- with iPhone support. So you could run legit OSX apps on the home server.
  • Full access to it as a regular Mac. If I want to install EyeTV and a tuner it would become a pretty slick networked DVR device.
So basically you're looking at something which is like a Mac Mini in terms or processing power with a larger case to accommodate more storage and a customized version of OSX Server to go along with it. Probably in the $800 to $2000 price range depending on storage.


So there's this utility, Syncopation, that is designed to sync multiple iTunes libraries. That's a major problem for me since I routinely use more than one Mac and iTunes is too ignorant to auto-monitor folders. Seems to be working well besides some crashing. Kinda slow but that's to be expected.

This is something Apple should do in OSX natively. Bonjour sorta does it but not really. I want to sync the actual data -- not just view what else is on my network at the time. Some sort of peer-to-peer home folder sharing would be ideal. You can do network home folders with OSX server but that's going to be overkill for most people unless Apple does a home server.

netatalk on Ubuntu

Good guide on how to get netatalk (AFTP) working on Ubuntu:

Very annoying that netatalk is still broken by default. Anytime I do an apt-get upgrade it stops working and I have to remember how I fixed it. Since I don't really have any Windows machines I don't see much reason in using SMB. The Samba team has done an amazing job but they cant change the fact that SMB/CFIS is just an ugly/slow protocol. I see an extra 10MB/sec using AFTP on the same hardware/network.

Exit Polls

Just a random thought, election day exit polling is going to be super fucked up this year. It was pretty bad in 2004 but with the huge increase in early voting it's going to be pretty much worthless this time around.

-In some battleground states a third of all votes will be made early. Combined with military votes and other absentee voting I could see the exit polling showing McCain up 15% OR Obama up 15% -- it's just going to be too inaccurate to make any real sense of.
-In 2004, for whatever reason, Kerry was hugely over represented in exit polling. A lot of Bush voters either didn't stop and answer and/or they lied because they were ashamed of voting for him. The racism angle might play out the same this year. I've noticed that racists are often really ashamed of their views and try very hard to over compensate when they feel like they are being scrutinized.

So anyway, no one should get too excited or too depressed over exit polls this year. I suppose, given the choice, I'd rather see Obama lagging in exit polling to motivate the get out of the vote effort. The first leaks tend to happen mid-day so you've got 4-6 hours of east coast voting and 6-9 hours of west coast voting. A big Obama surge in the exit polls would probably be really bad. I don't think it will come down to winning NV but I'd definitely like to seal the deal there too just in case.

Tuesday, October 28, 2008


Got the last remaining bits today (heatsinks) Assembly took longer than I expected. It was tricky getting the board properly supported in a non-CEB case. These copper heatsinks are very heavy and they were causing the board to sag. I had to prop up some plastic motherboard standoffs for a temporary fix. I definitely need a CEB case. That will have to wait till I get back from Boston.

Speed is pretty incredible. I just did a VisualHub encode on "Go Nuts" h264 on a 350MB file and it took about 5 minutes. Haven't had time to fuck around with Logic much. I turned up 4 channels, threw on a directional mixer & compressor and the CPU bar was barely moving. Friday the fake-band is coming over so we can do some serious testing.

These CPUs run incredibly cool. Even after a full load the heatsink is actually cool to the touch. I'm proud of my super ghetto fan. I tried using 80mm fans but they were just too loud. I took a 120mm fan and mounted it diagonally using two woodscrews that just kinda jammed inbetween the heatsink fins. I added a few bolts and a rubber washer to raise the fan up above the CPUs enough to force out the hot hair. (it's blowing down onto the CPUs) Totally ghetto but it works amazingly well.

I'm giving up on the idea of doing VMWare ESX at home for now. Just doesn't make much sense for me. It's a really neat platform but I have one to play around with at work now.

Speaking of which... I think we're pretty much all set with it. Rob figured out how to make VLAN trunking work. We're using an NV chipset and I suspect VMWARE doesn't see the PCI-X bridge so we can't easily add more network interfaces. We're going to try a PCI-E but, again, VMWARE probably won't see it. Not a huge problem but it's a limiting factor down the road to some degree.

OpenFiler + iSCSI is working well. Need to figure out HA eventually and, next year, we'll have $4k to spend on another SAN and another octo-core ESX box.


XBench: iSCSI

iSCSI on OpenFiler, on a u320 controller/drive, over GigE:
Results 13.04
System Info
Xbench Version 1.3
System Version 10.5.5 (9F33)
Physical RAM 3072 MB
Model MacBookPro4,1
Disk Test 13.04
Sequential 7.57
Uncached Write 51.02 31.33 MB/sec [4K blocks]
Uncached Write 47.33 26.78 MB/sec [256K blocks]
Uncached Read 2.20 0.64 MB/sec [4K blocks]
Uncached Read 30.68 15.42 MB/sec [256K blocks]
Random 47.08
Uncached Write 18.08 1.91 MB/sec [4K blocks]
Uncached Write 87.45 28.00 MB/sec [256K blocks]
Uncached Read 130.51 0.92 MB/sec [4K blocks]
Uncached Read 94.66 17.56 MB/sec [256K blocks]

That's with an MTU of 1500. I can't seem to turn on jumbo frames on the OpenFiler box. I would speculate that jumbo frames would help out with sequential reads/writes significantly. I think perhaps there's a BIOS setting to enable/disable jumbo support but it's too late to fuck around with it much more.

Monday, October 27, 2008

NAS things

So I got iSCSI working with the GlobalSAN iSCSI initiator. It wasn't entirely obvious that you had to "Map LUN" in OpenFiler to make things work. In fact, nothing in OpenFiler is very obvious. I'm not digging it so much. I think it would be OK if you were starting fresh but it's just not designed to pickup existing volumes. Everything has to be LVM based. I suspect FreeNAS is setup in a similar way. The installer kept hanging on my hardware so I don't know. I'm not too keen on using any software I have to fight with so it's not really an option.

iSCSI is definitely neat. I'm thinking of lots of different ways to use it already. If I actually want to use OpenFiler though I'm going to need to invest in some new storage and start fresh with LVM. still has those $50 500GB drivers so I'm thinking:

6x500GB in RAID5 = 2.5TB (maybe a hot spare too)

And I can setup the existing drives in RAID mirrors.

250GB for audio
300GB RAID stripe for Logic -- although I don't know if iSCSI is going to increase disk IO latency too much. Need to research that.

Plus whatever other left overs I have for TimeMachine backups, document backups, etc.

Redistribution of wealth

Here's why the "redistribution of wealth" attack doesn't work: Most people don't make over $250k a year. Most people feel like the rich cheat their way out of paying their fair share of taxes anyway. People making under $250k a year will get a tax cut. People who are really upset about taxes, especially the ones making more than $250k a year, were already voting for McCain. 95% of all small businesses don't make over $250k a year of taxable income. Cutting taxes for the poor & middle-class end up helping small business anyway.
Companies have been moving overseas and laundering money through off-shore accounts even while Bush was cutting their taxes. There's little or no proof more tax cuts for the wealthy would stop that trend.
The same attack was used against Bill Clinton and things worked out just swimmingly for the poor & middle-class.

Most Americans are smart enough to realize we're paying for 2 wars and a huge financial bailout. They understand someone has to pay the bill. People who got rich off oil futures, war profiteering, and cut throat lending are the ones we want to pay for it. Not us.

Sunday, October 26, 2008

The Final Week

State Level: Even more competitive states in the picture now. GA, IN, maybe even AZ. I would consider all 3 of those long shots at this point but it really wouldn't surprise me to see at least one go blue. The Obama campaign seems to be focusing on NC & VA right now after recently completing a final trip out west to sure up CO & NM. NV I think is still a very competitive race. This is another state that simply doesn't fall into line behind a more national political trend.

McCain continues to focus on PA, IA , FL & OH. No indication of movement in PA or IA. The race remains more or less tied in OH & FL and those are must-wins for McCain. Given the closeness of the race I do think it's possible McCain could still win both FL & OH -- in which case he would probably win NC, GA, IN, etc. (if undecideds or leaners break back towards McCain)

The difficultly for McCain, even if he wins OH & FL, is losing VA's 15EV. It would be absolutely crucial for him to fight back in VA or PA this week to win the election. Alternatively he could focus on winning WI, MO, NH & NV but that's a tough logistical battle to fight this late in the game. There just isn't enough time to be moving chess pieces all over the map.

National: Nothing new at all. McCain's attacks have been ineffective in terms of national polling. Probably too late in the game to stage another strong attack. He would need one huge, opinion changing event, to have a chance now. Too many states, not enough time for him.

Election Night: The results from VA & NC should give us a pretty good idea how this thing is going early in the night. By 10PM it will be pretty clear who won the election. If I had to predict I'd say Obama get to 270 before midnight.


You'd think with the names being so similar it would work. Apparently VMWARE won't support any regular ole' SATA controllers -- only SAS & SCSI.

I guess that forces my hand. I'll have to do a FreeNAS or OpenFiler NAS and a separate ESX server. I'll probably use a GigE crossover cable between the two boxes for a dedicated SAN backbone.

I think this is definitely a better way to do things but I have a deep hatred for artificial software limitations. I know my 3WARE card works in Linux. I know SATA is reliable enough for what I'm doing here. I should be able to make it work. It's my stuff, I decide to take the "risk"

Anyway, that means I need to buy that Lian Li case now. Probably won't get to finish this project up until after my trip to Boston now.


Random thoughts & notes about VMWare ESXi:

  • Installation is very straight forward although the installer itself is incredibly slow. It took a good 5-10 minutes to start it and the actual install took maybe 2 minutes at most.
  • VMWare Convertor will let you convert a standard workstation/player VM to Infrastructure and automatically upload it to your ESX server. Nice feature.
  • You can create a datastore to hold all your ISO images of operating systems so setting up new VMs through the Infrastructure client is super easy and quick.
  • Unfortunately most of these tools are Windows or Linux only. I don't think OSX Fusion can interface much, if at all, with ESX.
  • I can't seem to pass through a local drive for the VM to access natively. I think I'm doing something wrong but I'm not really sure. Perhaps it's just not detecting my 3Ware card in which case I'm going to need to build a separate NAS -- which is probably a good idea anyway.
  • The whole datastore concept makes a lot of sense. A big RAID10 setup would probably be ideal for this. Overkill for my personal needs of course but I think we'll end up doing that at work. Perhaps two OpenFiler boxes in HA mode, each with a RAID10.

Saturday, October 25, 2008

Improving Ubuntu 8.10's UI

Based on screenshots of Ubuntu 8.10 RC:

1) The root menu / start menu is horribly outdated. It's basically a less functional version of the Start Menu Microsoft shipped with Windows 95. It needs an integrated run/search box at minimum.

2) It would probably make sense to combine Applications/Places/System into one menu. It wouldn't preclude you from doing it separately if you wanted to -- there needs to be a panel app that offers an all-in-one setup and it should be the default.

3) Nautilus has way too much wasted space. The breadcrumb bar and toolbar should be smaller by default. In fact, the toolbar for all GTK apps should be smaller.

4) Part of the wasted space problem is fonts. The default GNOME fonts are just too big. They need to go down -1pt at least.

5) On post-install the user should be presented with at least a few different color/theme choices. The orange/brown thing really doesn't work. It's fine to include it but at least have an option of something different the first time a user logins in. Many people will never take the time to carefully customize themes. There should be several pre-built & tested color variations.

Other than these 5 suggestions I think they've reached a nice level of stability and consistency with the Ubuntu/GNOME UI. It doesn't need any major changes -- just some refinements.

Friday, October 24, 2008

The whole thing with McCain/Palin is enough is enough. I don't really think McCain would have been a bad President maybe 8 years ago. In fact we'd probably be much better off if he had beat Bush in the 2000 primaries. Although Gore's personality defects would have probably looked tam to McCain's so there's a very good chance Gore would have won by a big enough margin to escape the voter fraud in FL. Either way, McCain made himself unacceptable to voters by doing a few simple things:

1) His style of politics is just so dirty and divisive. Enough is enough. We just can't do this anymore. Calling people terrorists, socialists, etc it's non-sense. It's the worst type of politics

2) Palin is completely unacceptable. I'm not comfortable with her as VP. In fact, I'm not even comfortable with her as Gov. of Alaska and I don't even live there. I'm probably OK with her being Mayor of some small town in Alaska. Seems like a good job for her.

3) Despite what McCain says, you'd be hard pressed to find 5 relevant issues he disagrees with Bush on. Even the issues where he sorta disagrees (global warming) he basically advocates the same exact policies. For example, if he thinks global warming is caused by man why is he advocating that we use even more oil? Immigration is another issue he supposedly disagrees with Bush on but in reality the Republican party drove Bush far right on it just like they're doing to McCain. Gov. Bush had a far more liberal view of immigration than President Bush. So even on that issue they're basically the same.

4) Stop trying to divide America. There's no "real America" White people who live in rural areas are no more real than a black guy living in the ghetto. If you can't accept that you need to move to a country where the law explicitly allows people to be treated as lesser human beings. Saying "oh she didn't really mean it" isn't acceptable. You say words, they have meanings. Everyone knows that "real America" is a codeword for white, christian, rural voters. If you want to say it, say it... Palin wants to say "Obama hates white people" but she had to use codewords designed to entice the base.

5) They truly crossed the line with some of the hate this year. It's one thing to attack your opponent on policy, personality, qualification, intelligence, etc but when you start calling people terrorists you have failed some basic test of American decency and are instantly unqualified to be President. The extreme left does it too but we've never went so far as to have our candidates for office or official campaigns do it. That type of extreme fringe attacking has to stay at the fringe. When the candidates start talking about it it gives it a level of legitimacy that it simply doesn't deserve. A left wing commenter on a blog saying Bush is a nazi is quite a bit different from Palin saying Obama pals around with terrorists. There's a time and a place for those types of attacks -- they're part of the political process unfortunately but they have to be kept out of the mainstream or we're only a few steps away from political violence erupting in this country. Any candidate who is willing to do that is simply not qualified to be President of this country. They'd be better suited to run for President of Iran where extremist fringe politics is the norm.

6) America doesn't like angry politicians. It's hard to find any modern examples of a Presidential candidate who has won on a platform of anger. I think the only reason Kerry lost in 2004 was the fact he never managed to come off as anything more than someone who was really angry at George W. Bush. I fully accept that McCain hates Obama. His campaign has at least been successful in getting that message out there. OK. So what? Just being angry at someone doesn't make you qualified to be President.

7) It's kinda hard to win an election when you're using these huge sweeping attacks on a guy who is ahead of you in the polls. You're basically attacking the majority of the electorate at that point. If you say Obama pals around with terrorists then naturally that means anyone inclined to vote for Obama is a terrorist lover who hates America. Once voters start to lean towards one candidate those attacks suddenly become personal attacks on the voters themselves.

8) The radical right has a death grip on the Republican party. They are driving people away in record numbers. Democratic registrations, country wide, are up, independent registrations are up. Republican registrations are either flat, or down. They have allowed the extremists to turn their party into something resembling a radical Islamic fundamentalist organization. Do a find/replace on allah/god and you pretty much have the same thing. As I've said before, it wouldn't surprise me in the least if they actually do resort to violence. They've done it in the past with abortion doctor murders, the OKC bombing, etc. Out of all the mainstream political organizations in this country they are, by far, the most likely to commit a major act of domestic terrorism in the next 20 years.

Thursday, October 23, 2008

Is there any chance PA will flip? (2004 exit poll)

Short answer, no.

Fact #1: Obama's advantage with female voters is on-par with Kerry's support in 2004 at 54% -- 4% undecided or "other" To get the same level of support from female voters McCain would have to win the entire 4% of the uncommitted vote to get himself to 46% to match Bush 04.

Fact #2: Obama surges past Kerry's 04 support among men by 5%. 7% uncommitted. McCain would have to grab the entire 7% undecided and erode 3% from Obama's lead to match Bush 04

Fact #3: The most recent polls out of PA are under representing the black vote by as much as 3% Obama will win that 3% by an overwhelming margin and this is all assuming 08 turnout matches 04 turnout when in reality it will be much greater.

Fact #4: Obama is not drawing a significant amount of extra support among young voters in PA. He's presently on-par with Kerry 04 at 60% 18-29. So in other words, Obama is not relying on a flakey demographic to win the state. Obama does not need radically higher youth turnout to win. Again, McCain would have to win all 5% of the uncommitted voters to match Bush 04.

Fact #5: Obama is exceeding Kerry 04 in 65+ Yet again, McCain needs to win all 6% of the uncommitted voters and erode 1% from Obama to match Bush 04.

Fact #6: Obama has *double* the support among Republicans in PA as Kerry 04. McCain is grabbing an additional 3% of the Democratic vote in PA. Most likely angry Clinton supporters. One of the few bright spots in these numbers for McCain.

Fact #7: Only 12% of voters in 2004 made their choice in the proceeding week before the election. This suggests PA voters tend to make up their mind early. Only 10% made their choice in the month of October. 79% made their choice before October.

So in summary... Obama's advantages are across the board. There is no particular demographic McCain can target. McCain is facing large deficits from Bush 04 in almost every single demographic at a more or less equal rate. He would need an across the board surge of approximately 10-15% to win PA while there is a strong indication that PA voters are abnormally hard to sway in the closing 4 weeks of an election.

OpenFiler on VMWARE ESX

So I think I've finally got this straight...

File Server VM: OpenFiler -- doesn't do AFTP but I'm really interested in trying out iSCSI. Why? Because I can. Plus, CFIS is just an ugly protocol. Way too much overhead. I'm not sure AFTP is much better. Lots of wasted packets. I'll probably need a CFIS share setup as a fallback though.

Downloading VM: I think I might consolidate this with my OSX Server VM. With screen sharing I can use Transmission which is about 1004.5% easier to use than rtorrent. Hella runs on anything so that's not a concern. I can throw the Logic node on there too. (not that I really NEED a Logic node with an octo-core system but hey, why not) I'm also interested in trying out networkable spotlight. I have 3 Macs so it's probably not a bad idea. Maybe home folder re-direction too?

And I don't think I need a personal Exchange. Too much trouble, nothing really worth saving.

Hazel for OSX

Somehow I never heard of this app before but it's really fantastic. Basically it just lets you monitor a folder and setup rules using a very flexible criteria matching system. (kinda like SmartPlaylists in iTunes)

So for example, I just setup my downloads folder to automatically sort PDF files into a sub-folder called Documents, DMG/EXE/ISO files into another one called Software, and to delete incomplete downloads after a week.

The most interesting part of this app is the power to run AppleScripts & shell scripts. That gives it unlimited power. I'm trying to figure out how to make a shell script that automatically encodes new video files into an iPhone/ATV friendly h264 using ffmpeg and then send a growl notification when it's done.

So many great OSX apps out there these days.

TIME article on Obama

Even more remarkable, Obama has made race — that perennial, gaping American wound — an afterthought. He has done this by introducing a quality to American politics that we haven't seen in quite some time: maturity. He is undoubtedly as ego-driven as everyone else seeking the highest office — perhaps more so, given his race, his name and his lack of experience. But he has not been childishly egomaniacal, in contrast to our recent baby-boomer Presidents — or petulant, in contrast to his opponent. He does not seem needy. He seems a grown-up, in a nation that badly needs some adult supervision.,8599,1853025-1,00.html

OH, 2004 vs. 2008

For whatever reason I really enjoy reading some exit polling.

Here's why OH is still competitive when the national trend has so clearly broken for Obama:

1) Black voter turnout in 2004 in OH was high (10%, 3% increase over 2000) Obama is seeing a large bounce from black voters in some of the other battleground states that is simply not realistic in OH given the already high turnout. Say most likely there is an additional 1% turnout increase vs. the 3-5% you will see in some other battlegrounds.

2) More higher income voters. ($50k+) Bush won them 58/42 in 04.

3) The Republican Get-Out-The-Vote effort in 04 was quite good. 15% first time voters, 46% went to Bush, 54% to Kerry. That's still a big advantage for Kerry but it's probably going to be double that in 08 for Obama.

4) Strong support in southern OH.

So OH is a competitive state. It will really depend on a surge in Democratic turnout along with smaller increases across the board. For example, there are more Latino voters in OH today than in 2004. Obama is doing better than Kerry among Liberals. (14% of them voted for Bush in 04, believe it or not)

Another interesting factor is the state of the race in southern OH where McCain has very strong support. A recent poll showed Obama doing 2% better in KY than Kerry did. That's important because southern OH's demographic is very close to KY. If Obama grabs even 2% of that vote he would hurt McCain's chances at carrying the state.

The most important thing here is the Obama campaign never focused on OH as a must-win. Their strategy from day one has been to win without OH & FL but to stay competitive enough in those states to keep McCain's resources busy. So far, the strategy has worked perfectly.

Wednesday, October 22, 2008

apt for OSX86

I think it would be useful for the community to develop something along the lines of apt for updating OSX, installing drivers, etc.

Right now it's just a little too confusing trying to hunt down forum topics to find the most up-to-date methods of doing things and any required files. These methods also depend greatly on minor details that people always overlook.

It would have to know what type of platform you're using (AMD/Intel/NV/ATI) and also what distribution you're running along.

So for example, apt-get install nvidia might prompt you for your graphics card device ID and, if possible, automatically install or at least tell you what your options are (EFI, standalone injector, modified KEXT's, etc)

It would also be very helpful for doing OS updates. Also for installing OSX86 utils (kexthelper, efi studio, etc)

Anyway I suppose it hasn't been done for the same reason I won't do it, it's time consuming and really only solves problems for people who can't be bothered to read & follow instructions. Still, it's a good idea for someone to work on since it's pretty easy to pull off.

New Case for "MacPro"

I found some good high res internal pics and it looks like a CEB motherboard should fit with a few modifications.

1) The rail that suspends the 120mm fans will have to be moved back towards the drive bays to accommodate the extra 1" of motherboard.

2) I'll probably have to drill an additional row of holes to support the edge of the motherboard. There will be a solid 1" overlap from the last screw hole and this happens to be the heaviest part of the board too. Plastic standoffs will probably work OK for this.

That should be about it. Luckily CEB isn't too huge. If I had bought the 5100x board I was looking at I would probably need a legitimate E-ATX case.

I also found a case that looks somewhat similar to the old NeXT Cube in terms of front panel design. That might be fun too. I'd have to find a vintage NeXT case badge.

Tuesday, October 21, 2008

McCain & PA

It's pretty clear McCain's last gasp is trying to flip PA. It's the only way the math works for him at this point. Under this scenario he also needs to win OH, FL and (at least) NV or another battle ground with 5+ electoral votes.

Obama snuck up on them so quickly in VA and has such a huge advantage on the ground that it's no longer realistic for the McCain campaign to organize and turn it around. There's also mathematical reasons for giving up on VA. 13EV simply isn't enough to push McCain to 270. He would need to win CO, NM, NV and MO also and there's virtually no statistical chance of that happening at this point. Chances are, at worst, Obama wins 2 out of the 4. (CO & NM most likely) Here's another math problem for McCain/Palin: If Obama wins VA, NM/CO/NV/MO are still not enough. In fact, even if you give them 1EV from Maine it's still not enough. (3EV short)

Here's where things start to look impossible for McCain. He needs to win PA, a state that broke heavily towards Obama 3 weeks ago and has shown no signs of turning back, and here's where it gets hard: He STILL has to find 13 more EV to win it.

Their last hope is to win 13 EV out of the following states: NV, NM, CO, MO. A total of 30 EV. Unfortunately for McCain, again with the math, it's not really 13EV. It's actually 16EV because no combination of these states gives you 13EV. The closest you can get is 16 (MO+NV) McCain might, as an extreme longshot, be able to pull 1EV out of Maine however MO + 1EV still isn't enough (1EV short guys, sorry)

So that's where we stand today: McCain must carry PA, MO and NV to win this election.

NV is winnable. MO is hard to figure. Not a lot of polling there in the last few days. One outliner suggests McCain is doing better there than previously polled. I guess the logic there is if PA goes red, MO would follow (same demographic principals) However this strategy fails to anticipate Obama's strong support in IA. This is another state Obama has run away with over the last month. Whatever groundswell of support allowed Obama to surge in IA is surely effecting MO too. Plus of course, flipping PA this late in the game given Obama's surge there is also incredibly hard.

Oh and this assumes Obama doesn't just win one of the following states: OH, FL, IN, GA, NC. If he wins even one of those it doesn't matter what happens in PA/MO.

It's not over yet. It's close though. Look at McCain's numbers in PA, MO, and NV at the end of the week. That should tell us if there's even the most remote chance he can make a play at it.


Interesting GoogleTalk on LLVM. It's a big part of Apple's future, already being used on the iPhone and the core of the improvements in 10.6 The whole GrandCentral thing makes a lot more sense to me now. I'm hoping to dig a little deeper on it but there's not much information out there yet. I'm really curious if GrandCentral will be able to supplement/replace XGrid as a distributed computing platform that is (mostly) transparent to the developer? Clearly it would never be as efficient as a carefully crafted native distributed clustering code but it would be a start. We have all this CPU & GPU power going idle most of the time. Whoever figures out how to harness it is going to have a huge head start. There's really no doubt the future of computing is increased core density and more general purpose use of the GPU. The benchmarks on SETI/Folding running on GPU vs. CPU is a testament to that.

When you see stuff like this you realize how far behind both Windows & Linux are. Apple seems to be the only company really pushing innovation in desktop operating systems these days. Windows is stuck in the past, Linux is stuck trying to copy Windows.

Nice Image

Lovely image... need to get a poster + frame:

TIPP Daily

Obama 46.9 (46.7)
McCain 40.9 (41.4)

Definitely not inline with Pew. seems to think Obama's cell-phone advantage is 2-3% so 6% + 3% at least gets its it consistent with Gallup Expanded and close to the MOE for the Pew poll.

GWU had a national daily out today, Obama +1 So in one day the accuracy of pollsters is a MOE of at least 14% and maybe as high as 18% Cell phones, turnout, and racism must be making this a very difficult election to poll. By this point in both the 2000 & 2004 races the polls were more or less agreeing with each other within their MOE.

Death of the Mini?

The rumor is Apple Stores have been told not to expect anymore shipments of the Mini. So that means either Apple is killing the Mini or planning a major revision. Hard to say since the Mini isn't a big seller for Apple. It seems suspicious they would stop shipments 3 months before the replacement model but if sales were low it would certainly make more sense than building a bunch of new Minis right now. Apple has also shown a willingness to take a short term hit when introducing new products. The original iPhone was unavailable at least a few weeks before the 3G launched. Could be the same situation here.

I guess the other possibilities:

1) The Mini is getting a total redesign -- perhaps more in the flavor of a delux AppleTV. If the rumors are true about Apple introducing TVs running ATV software it would suggest they are looking to expand their set top / living room presence. It would certainly make sense to offer the existing AppleTV at a much lower price point. $149, $99 even. Subsidize it with an iTunes gift card at very least. If Apple ever gets an iTunes Subscription model going the AppleTV should be $99 or less -- free with a 2 or 3 year subscription.

2) The Mini is turning into a mid-range desktop Mac that everyone wants. Using the same industridal design of the new MacBooks with the usability of the original G4 tower would be quite nice. NVIDIA chipset + 2xPCI-E 16x. Dual core or Quad core options. If they go this route they should introduce a 20" LED cinema display too and offer a decent discount on bundled computer + LCD configurations.

I don't believe a mid-range Mac tower would cut into sales of the MacPro. People who are buying the MacPro truly need every ounce of power. They won't be downgrading to a quad core when they can get an octo core MacPro. The price delta would likely be in the same range as the MacBook vs. MacBook Pro and there's absolutely no indication to this point that MBP sales have been hurt by the lower price point / power of the MB.

Pew Update

Registered Voters
Obama 52, McCain 38

Likely Voters
Obama 53, McCain 39

Most importantly sample size is over 2,000. Very large for a national poll. I believe that's double, even quadruple, a lot of the other nationals. Sample size directly increases accuracy.

As I covered before I consider Pew among the best national pollsters. They've been in the business for a long time and don't seem to have a partisan axe to grind. When you look at the state polls it's hard to imagine Obama only being up 4 or 5% national. How can he be competitive in NC, IN, VA, etc? I can't imagine those previously red states are somehow unique on a national level. Clearly Obama must have very strong support nationwide to be pulling ahead or even in those states. +10 to 14 would make a lot of sense given the local state margins.

Of course then you have OH & FL. Perhaps they just don't follow the national trends. FL especially is a very artifical demographic. Lots of snowbirds, retired folks, etc. Their results just may not be in line with national trends. OH is a bit harder to figure. I would think Obama should be seeing a solid lead there right now given the situation in other battle grounds and national polls. I'm not too familar with the demographics of OH though... perhaps the southern part of the state is really breaking overwhelmingly towards McCain and showing strong signs of likely turnout. That would explain it.

Looking at the racial balance in this poll doesn't seem to make a lot of sense. Obama's lead might, in fact, be bigger:

Pew 10-21:
White 81%
Black 10%
Hispanic/Other 9%

Actual 2004 demographics:
White 77%
Black 12%
Hispanic 8%
Other 3%

National polling 10-21

Looks like the national numbers are falling back into line again. It'll be interesting to see if the Powell endorsement is having a real effect or we're just seeing the typical weekend vs. weekday polling swings. It's probably a good thing these weekend numbers shrink Obama's lead -- it'll keep people on the ball and prepared to actually vote. It's too easy to say "he's going to win anyway" and blow it off.

New polling out of IN is almost too good to be true. PPP has Obama up 2. I think these numbers are probably closer to being an outliner but there's no dobut Obama is surging in IN. That's probably going to help in north-western OH also. Southern OH is a lost cause -- those people would vote for Hitler over a black Democrat. McCain can have them.

NC looks solid too. I kinda doubt we'll win it but it bodes well for VA. If Obama wins VA it's over unless McCain can flip PA. Given how many resources McCain is putting into PA I'm more than a little worried he knows something we don't. Perhaps the racist country people are giong to come out in droves to make sure a black man isn't President. I guess we won't know what the hell McCain's strategy is there until after the election is over. I do think PA is his best bet to flip a blue state so perhaps he's just fighting there because there are no better options. Given how the other battleground states are going flipping PA might not even be enough for him anyway.

If we can finish this week with some strong numbers I think it becomes virtually impossible for McCain to turn it around. He needed to start moving the polls 2 weeks ago. Right now we're just seeing the typical polling/MOE bounces. There's really no concrete evidence to suggest McCain is making any headway.

Conceptual view of a radically new Finder

I thought it'd be fun to explore how this might work.

First of all, all the computers on your network need to talk to each other and share their various home folders or other user defined sharing locations. Very easy to do on the Mac side with Bonjour -- integration with Windows shares would be more tricky and probably require some Apple software on each machine to make the process easier. For security reasons we'd have to worry about authentication across machines. A simple wizard that went through every machine found in Bonjour and asked for credentials would solve this problem.

For data you would have a few pre-defined portals. Video, Music, Documents, etc. It should be easy enough to match these locations on a Windows machine (My Music, My Documents, etc)

Once a portal is created you check which machines (local & cloud) that you want to include. Color labels or icon branding would identify where the file actually was. A file unreachable would obviously be grayed out or otherwise noted as being offline.

So now you'd open your Music portal. You'd be presented with an interface similar to a stripped down iTunes. You could browse by album, artist, or go into CoverFlow view. This would pool all the music, on all your machines, and present them as one large collection. A context menu option would exist along the lines of 'Reveal Location' where you could open the actual share the file is located on.

Other context menu options might include "Copy local" - "Copy to iDisk" - "Sync"

Sync would copy the file to every machine and monitor the file for changes.

Copy local or, make available offline, would copy the file to a specific machine.

This whole system would be greatly enhanced by building into OSX a good, easy to setup, VPN solution. So when you're at work with your laptop your portals still function. Location awareness (either by GPS or manual selection) could offer an extra level of control. For example, when I'm at the office my extensive collection of marginally erotic 1920's pornography probably shouldn't be part of my portal.

The Pictures portal would be very similar to iPhoto -- if you wanted to edit the phones or do any serious work iPhoto itself would open.

If multiple versions of the same file exist locally and remote, portal prefers the local resource after checking for consistency.

Portals would replace the current open/save dialog. The dialogs would need to present the user with enough information to make this thing practical. For example, opening a 4GB HD iTunes rental over 3G probably isn't a good idea.

The whole concept would rely heavily on good peer-to-peer VPN (Hamachi style?) and iDisk. With that in mind, iDisk should be part of OSX. Perhaps 1 or 2GB for free with a reasonably priced upgrade option. I would personally be willing to spend quite a lot for cloud storage if it was integrated into everything I did seamlessly. This would be an excellent business model for Apple going forward.

Integration with future iPods and iPhones is obvious.

Integration with Windows should be a best effort sort of thing... or you know what? Just forget it entirely. Make it a Mac only feature. Trying to deal with Microsoft on something like this would be a nightmare. I am doubtful it would ever work 100% so declaring it as a 100% OSX exclusive feature is far more honest and sets the expectations for the end user.

It's an interesting idea. It has its share of problems but I think it's do-able and we're long overdue for a more modern, cloud aware, way of dealing with data. The emergence of wifi & cell dat, along with faster residential broadband, makes it a reality Apple could pull off today.

20 ways to improve Finder in 10.6

So apparently Apple is finally making Finder a 100% native cocoa app. I'm not sure what took so long. Finder in 10.5 is much improved over 10.4 in terms of stability in performance but the new feature(s) are underwhelming. CoverFlow is useful sometimes but I feel like it needs a lot of refinement. For example, creating thumbnails is just too slow. I can manually find the file quicker than it can make a thumbnail for me.

If they're re-writing Finder it's probably safe to bet they're looking at revamping it also. I'm a fan of Finder's simplicity and usability. For any serious file management I use Terminal or Path Finder. I don't want to see them make Finder too cluttered but, at the same time, it feels like a very dumb app. I find myself repeating the same clicks over and over again. So for 10.6 I'd like to see:

1) More view options -- especially a view option that takes into account the type of data I am looking at. CoverFlow is a good start but there's a ton of room for improvement. For example, when I'm looking at a folder full of MP3s I should be able to see the ID3 tags. Windows Explorer has had this feature since Windows 2000 IIRC.

2) An option to automatically zoom the window size depending on folder contents. This is a feature I've used in rox-filer (I think..?) on Linux. It's quite handy. I find myself hitting the zoom button way too often.

3) SpotLight needs to default to the current folder I am in. It makes no sense to search the entire computer first. Perhaps it could be done with a prefix. For instance, if I type in _filename it knows to search the current folder first.

4) A highlighting find-as-you-type feature ala Safari where the unmatched items get grayed out and your search term gets highlighted.

5) TABS. I just have too many Finder windows open sometimes. Make sure they are movable and detachable like Safari.

6) Genius for Finder: If I'm always working on file A, B and C together a sidebar should show me them. If I'm always opening another folder when I work on file A, B and C, show it in the sidebar too. All the data is there -- it just needs to be parsed.

7) Icon view thumbnails. CoverFlow is nice but probably the #1 thing I miss about Explorer or Nautilus are the icon based thumbnails. It's just a more efficient use of screen space.

8) Intelligent font/icon sizes. If I'm looking at a folder with 5 items why not make the text/icon bigger? if I'm looking at a folder with 1000 items make them smaller.

9) Copy & Paste. It's about time guys. Personally I am able to cope with copy and manual deleting but the average Windows switcher is probably really put out by the lack of this feature. It's harmless to add, so why not?

10) Better Spotlight for network shares. This isn't exactly a Finder issue per say but it's something we need. I know it can be done with OSX Server but that's overkill.

11) More toolbar options. They don't have to be on by default but it would be nice to customize them. Mount network share, browse network, copy/paste, etc.

12) Stop doing that thing where sometimes a Finder window open that is rooted in nowhere. You can't go back, you can't turn on column view to see where the folder is on your file system. It's confusing and it wastes time.

13) Add a toolbar button for application expose so I can quickly see all my Finer windows.

14) Rectangle drawing selection in list-view mode.

15) Get Info sidebar /w hotkey.

16) Customizable shortcut keys. Hide it away somewhere in System Preferences so the average user who doesn't care about customizing them never has to bother with it.

17) I like how Finder remembers what items I have expanded in list view but sometimes it's beyond annoying. I'm not sure how to fix it but there has to be a solution.

18) Warn me if I'm about to accidentally open a gazillion files. I do this probably at least once a week. It's a little on the nagging side but I kinda doubt many people do it often on purpose so why not?

19) "Install" option on DMG packages. It's kinda cute to drag the icon into your Apps folder and all but an item in the context menu to do it for me would be pretty slick. Install the app, automatically unmount the DMG, done.

20) QuickView should keep playing audio & video when Finder is in the background. I really like the simple UI for viewing movies & audio and I would like to just use it as my primary media player most of the time. I rarely need more features than play, pause, seek. As things stand right now as soon as Finder is not the foreground app it stops. It's a little contrary to the logic behind QuickView, I admit, but I think it makes it a more valuable feature for people.

And, of course, option 21) Blow it up. Totally rethink the way Finder works. Free us from this archaic file/directory structure we've been dealing with for decades. I don't really care where my files are most of the time. In fact, I would truly love a way to mask that information for most of my day-to-day tasks. I would love to just have some virtual concept of "OK, here's all your music sortable by album/artist/etc with artwork and all your other meta data" or "OK, here's every GarageBand project you've worked on with previews." Maybe one file is on my iMac, another is on my MacPro, maybe another is on iDisk. Who cares where it is? I tend to breakdown my usage of Finder for viewing only. Serious file management requires Terminal or PathFinder. It might make more sense for Apple to extend this concept. Finder should be a portal to your data -- not a file management tool. I would highly suggest Apple purchase the company who makes PathFinder and integrate it into OSX as the advanced file management tool while Finder become simply a way to access your data in a passive way.

Sunday, October 19, 2008

u320 is still pretty fucking fast

Weird thing is my XBench benchmarks show it being slower for reads/writes. I'm guessing the seek time must make up for it because the drive definitely feels faster. I'm working with a gigantic Logic project and it's very noticeable. Should be even faster when I get it into a PCI-X slot.

Powell endorsement

I don't typically think endorsements make much difference in elections but the timing of Powell's endorsement, combined with general distrust & disillusionment with the Republican party, might make a real difference this time. I think a lot of moderate Republicans are not happy with their party right now and look at someone like Powell as sharing their values & concerns. I see this type of Republican as the typical 80's Reagan Republican who is getting hit especially hard by the financial crisis who perhaps doesn't have any strong social issues keeping them tied to the Republican party.

Powell's endorsement is also very good for Obama's military and foreign policy credentials. Anyone who was on the fence about that should be reassured now that a man who was instrumental in the last 20 years of US policy is on-board. My rule is if you have a military doctrine named after you, you're opinion on military issues is worth listening to.

There's also the race angle... the radical right trying to make this endorsement a racial issue are making themselves look desperate, out of touch, and pathetic. There's plenty of good reasons for Republicans to support Obama. It doesn't have to be a racial issue. Anyone who tries to make it one is showing how brainwashed they truly are. There's quite a collection of white Republicans endorsing Obama too -- including William F Buckley's son. If you've got white people from the National Review supporting you I think it's safe to say it's not a racial issue.

From a media stand point Powell will be the dominate story tomorrow, perhaps into mid-week depending how far the Republicans want to go with the race baiting. If they really want to go down that road there will be a media backlash over it. That will give this story legs all week long.

Bootable SCSI + OSX

So System Profiler says I can boot off this card. Once this drive finishes formatting sometime right before the 2010 congressional election I will give it a shot. I don't think I will actually boot off it though -- I'm just curious if it will work. It's probably more reliable to keep my boot drive on SATA. I can also disable SpotLight indexing on the Logic scratch drive and I won't have to worry about any other disk IO hitting the drive at weird times.


So this u320 card I got works perfectly. This 300GB drive I have had some weird proprietary formatting though so I think it needs a low level format. It's going to take a while to do that. I started it once and 24 hours later I think I was at 8% complete. It's also possible the drive is just borked. If so, I'll probably just buy some SAS drives instead. The only reason I went with SCSI is the fact I already had this drive and the controller was $30 so... it was the cheapest option. Plan B is a pair of 15k SAS drives. Should be enough disk IO to keep up with Logic.

New TIPP numbers

In 2004 the two most accurate pollsters were TIPP & Pew Research. Both were within half a percent of the actual national popular vote.

Obama - 47.2%
McCain - 39.8%
Unsure - 13.0%

Obama - 49%
McCain - 42%

The Pew numbers are a bit older (10/15) however the same margin exists. The 13% undecided in the TIPP poll seems too high this late in the game. Perhaps TIPP accounts for some percentage of weak/leaning support as undecided voters. If that's the case McCain would need to win these undecided voters by a 60% margin to tie the national numbers.

The outliner today seems to be Zogby -- showing a much tighter race. In 2004 Zogby was off by about 3% showing a tied race when in reality George W Bush won by 2.4% national. So based on that data, Zogby underestimated the winner by 3%. 3% + 3% gets us close to the margins TIPP & Pew are showing. Gallup used a traditional likely voter model in 2004 and showed basically the same results as Zogby and the same under estimation of the front runner. -3% Gallup's numbers this year also line-up with Zogby in the traditional model however they now also offer an expanded model that is in line with TIPP & Pew.

So what does this all say? I don't know. I think by all accounts the race is stable however the likely voter models are a huge X factor. No one really knows who is going to show up and vote. I'm sticking with TIPP & Pew for the most accurate numbers because I feel like neither of them have an axe to grind. The big pollsters, being paid by the mainstream media, have less incentive to be accurate and more incentive to be entertaining. For conservatives it's fun to think it's a 3% race. That's almost margin of error. For liberals, it's fun to think it's a 10%+ race. That's a landslide.

With 2 weeks to go I expect the national numbers to stay mostly the same. We've seen no significant movement in the last 2 weeks throughout the debates. There are no more national events where voters can compare candidates side-by-side. The holiday season & baseball playoffs begin to effect the Presidental race in the last 2 weeks. People are less interested in following the day-to-day campaign news. It's less of a topic of conversation.

The only significant change I expect to see would be the last week of polling before the election. This is the point where voters will have mostly made up their minds. It will also include Obama's nationally televised campaign event.

Obama's tax plan

The Republicans are truly regressing. First Obama was a terrorist, then he was a lib-e-ral, and apparently now he's a communist. If we regress any further into the past they're going to have to claim he's a southern gentleman who wants to bring back slavery. That might be a tough sell. Maybe they can go back even further and call him a Redcoat sympathizer. He's always palling around with British monarchs and whatnot.

It always comes back to taxes with them. Specifically taxes for the rich although for PR purposes they sell it as "tax breaks for small business owners" since that's a viable populist spin on things. If they ran on a platform of "tax breaks for the rich" they probably wouldn't get many votes so they throw you a couple hundred dollars which you promptly blow on nothing. (most likely sending a good percentage of those dollars to other countries who actually produce the items you're buying)

Anyway... The truth about business taxes are, on paper, they are quite high. Why is that? Because there probably isn't a single company that actually pays 35% taxes. The tax rate has gone up over the years because the companies have found new and better ways to avoid paying taxes. Some legal, some not. We just want them to pay their fair share.

Under US law corporations have most of the rights of an individual. As an individual, if I had the money for a really good tax man, I could probably skip out on paying most of my taxes too. The problem is, for me, it makes no sense to try to skirt around the tax code because I'd have to pay someone a lot of money to do it for me. More money than I would save. So I just accept it and pay my fair share based on whatever the IRS tells me I have to pay. A large company on the other hand can hire a whole team of tax experts, skip out on paying their fair share, and save tons of money in the process. Is that fair? I don't think so. I don't think most normal Americans who aren't rich would think it's fair either.

McCain cited some fact that we have one of the highest business tax rates in the world. Technically it's true but in reality the amount of our GDP that goes to taxes is far lower than the countries he's comparing us to. Obama's plan raises the paper rate on businesses slightly and you know what? They won't pay most of it anyway but if every business pays 1% more that's quite a lot of money. (of course you have to ask for 5% on paper to get 1% in reality) The Republicans like to say this is "redistribution of wealth" however they happily take advantage of this country and its resources. Why is it unreasonable to ask them to carry their weight?

It's not like these US companies aren't getting something out of the deal. The US government does a lot of work for business -- trade negotiations for example. Just look at NAFTA, the trade in-balance with China, the government's casual support of human rights when dealing with countries we do business with, etc, etc, etc. We waged a war in Iraq to make sure their precious 18-wheelers can stay full.

Intellectual property law? We have this giant infrastructure of red tape to protect these companies. It's a cabinet level position now. Our law enforcement agencies have to deal with it everyday. Our judicial system has to deal with it. Think about how much money is spent every year on protecting IP. Who pays for it? ME? Why? I don't care. I don't have any IP to protect.

There are countless other examples but in an effort to stay on topic I won't get into it.

So why shouldn't they pay their taxes like I have to? I'm willing to bet the typical company gets more special treatment and benefits from the government than I do.

McCain was really hot for the "redistribution of wealth" thing in the last debate. It's true too. The rich have got richer and the poor have got poorer. That is an indisputable fact. Why is that? Did the middle-class suddenly stop working hard? Did a bunch of people quit their jobs? No. What happened was big business pushed them harder and harder so they could make more money and, incidentally, not pay their proportion of taxes on it. It's a correction, not an increase. The poor & middle-class have been carrying their weight the whole time. We grumble about it but we do it because we think it's dishonest to skip out on the bill. It's like going to a restaurant with your friends and everyone orders the most expensive thing on the menu but when the check comes they're pissed off. The Republicans knew they were spending more than we could afford. Now they're pissed off we want to pay for it. It's the most irresponsible thing I've ever seen.

And, let the record show, we wouldn't need a tax correction if we didn't have 8 years of unprecedented spending by the Republican party while they cut taxes at the same time. It's just basic kitchen table math. If a family spends more than they earn it will eventually catch up with them. The mortgage crisis is just a nice little miniature version of what this country has been doing for the last 30 years (with a brief 8 year interval where Clinton created budget surpluses)

The next time a Republican suggests the Obama plan is socialism or communism I wish someone would stop them and tell them the truth. McCain's plan puts us closer to socialism than Obama's does. McCain's plan just pushes the poor & middle-class further down to help big business while big business gets richer.

When the gap between the poor & rich becomes so great that the poor no longer have any incentive to be productive, law abiding, citizens you are creating an environment ripe for revolution. We were at that point in the 30s. We'll be there again if McCain somehow wins this election. If any of these Republicans think Obama is a socialist they're in for a shock when they see what will happen down the road if we continue these trickle down policies. Obama will look like William F Buckley compared to the revolution they will create down the road.

The Republicans just need to accept this as a correction. Eventually the Republicans will get back into power and continue their policies. It's healthy for the country to have this cycle of pro-business, pro-labour. The cycle has been thrown off in the last 30 years and we're suffering because of it.

Saturday, October 18, 2008

Finally got my 16 channel audio input to work more or less. Still the occassional peaking out. Not really sure why since none of the audio levels are very hot. At least it's minor enough to EQ out now. 

Friday, October 17, 2008

Pay to Learn — Los Angeles philanthropist Eli Broad has probably never met Soledad Moya, an eighth-grader at Middle School 302 in the South Bronx. But both are big believers in an approach that has people wringing their hands and wagging their fingers: paying students to perform on standardized tests

I like this idea a lot. I think it could work on a much larger scale. Nothing motivates people as much as money. Learning for some hypothetical goal down the road is just too abstract for a lot of kids. They want instant gratification. Let's face it, public education is nothing more than training people how to be good worker bees. Why not give them a real world application of it?

We're already throwing a huge amount of money into public schools and it doesn't seem to be working because the problem is motivation of students. No amount of technology or gimmicks can get around the fact that many kids just don't want to learn. They tend to bring others down with them too. I remember being in school and having trouble focusing on anything because of the classroom distractions.

So basically the kid gets $X for passing a test on a sliding scale.
Passing a term would be an additional $Y

I guess the real question is how much money is involved and how do you pay for it but I'm a tax & spend liberal so that's an easy answer for me.

Ahhh.... so THAT'S why....

WASHINGTON (AP) -- Republican presidential candidate John McCain's wife, Cindy, is reporting $4.2 million in income for 2007 in newly released tax returns.

Ah ha... so that's why John McCain's definition of rich starts at $5M a year. See, John is just a normal middle class guy like you or I. John the Senator. Fighting hard to protect people like you.