A Little Ludwig Goes A Long Way

My grandfather once told me “The day you stop learning is the day you start dying.” I’ve had a lifelong commitment to education and I am still learning every day. There is so much going on in education, the choices are broader every day, with so many efforts to increase access and lower costs. Some things I’ve been learning about:

• played around this weekend with Apple’s new ibook publisher — Tons of coverage of the event announcing this week, see for instance http://techcrunch.com/2012/01/19/apple-textbook-event/. The goal is noble — allow millions of people to create textbooks, targeting the iPad of course, and dramatically cut the price of textbooks, and the carrying weight of textbooks. The tool works although it is a little buggy yet. I made a first textbook — basically i poured all the portfolio company summaries from the ignition partners website into a textbook format (a tool that would automatically pour CMS content into a textbook would be handy). These textbooks are really just another form of app for the iPad with a dev tool that is substantially friendlier to use than Xcode. If you can author a powerpoint presentation, you can author a textbook. There is nothing super revolutionary about the resultant products but this is a good step towards electronic textbooks.
• signed up for a course at udacity.com — We believe university-level education can be both high quality and low cost. Using the economics of the Internet, we’ve connected some of the greatest teachers to hundreds of thousands of students in almost every country on Earth. Know Labs was founded by three roboticists who believed much of the educational value of their university classes could be offered online for very low cost. A few weeks later, over 160,000 students in more than 190 countries enrolled in our first class, “Introduction to Artificial Intelligence.” The class was twice profiled by the New York Times and also by other news media. Now we’re a growing team of educators and engineers, on a mission to change the future of education.
• thinking about taking a course at Digipen as well. They’ve done great work, the team for Portal came out of Digipen.
• at Wolf’s advice, learning about the Dalton research group at the UW. A traditional university setting but exciting content.

My brain’s a little tired but excited about the opportunities!

• Trello. I really want to like this, the simple notecard interface is nice. But a little too structured for me, I’d prefer basically free form notes and less database-y feel. And I need an iPad/iPhone app, an iPad version of this could be awesome. Maybe I just really want a little better organization tools in Evernote, I don’t use the Evernote folders and tags much, I don’t really understand when to use tags and when to use folders.
• Picscatter. Great way to create a Facebook timeline header picture.
• The Tivo iPad app seems to work very well. Way easier to use than the onscreen guide and tivo remote. I’ve also used the xfinity/comcast app which is not surprisingly a little clunkier. It is sad how marginalized Tivo has become tho, they really overplayed their hand. Tivo doesn’t seem to have created a win/win partnership opportunity with cable/satellite providers and so they have all created and pushed their own crappy DVRs. I am sure Apple has learned from their iPhone experience and Tivo’s missteps, I would expect Apple to create upside opportunity for their partners and to have much greater success than Tivo.

Call of Duty: Modern Warfare sets first day sales record. What stunning numbers. I don’t think Skyrim will do the same but a huge week for online entertainment. There is clearly huge demand for great entertainment content, pretty much insatiable demand. And why not, these games give hours and hours of entertainment, the per-hour price is super low.

• PlugBug. Ordered two.
• iZON. Sleek look, nice looking software, but some bugginess. When it all gets ironed out will be a fabulous product.
• Orchestra. Nice todo list with voice input (might be rendered irrelevant by Siri but I’m not 4s yet). Thanks chris
• Batch. An Ignition investment. Nice photo sharing tool.
• Exoplanet. If you don’t love this, you have to turn in your nerd merit badge.
• Drawtop. Haven’t ordered yet but the simplicity of converting an unused flat surface into a whiteboard is so appealing.
• Stepping into photos. No software to buy here, but I would buy in a second if there was something reasonably priced that would let me do this.
• Lost Crates. Love the idea. Not sure it would work well in practice.
• Voxatron. What a nice looking indie game. Doesn’t quite suck me in the way Minecraft or Darwinia did, but excellent.
• Findings. Thanks @crashdev, this should be just part of the Kindle experience.

November is shaping up to be quite the race in the Big10, I have to say, the addition of divisions and a championship game have created a great new dynamic (and makes me rethink my objection to a playoff, hmmm.)

Many sites have written about all the permutations of possible teams in the championship game, and the surprising fact that OSU is not out of it, and in fact has a very good shot at making the game. Being a bit of a nerd, I decided to play around with a Monte Carlo simulation of the race to the championship game.

I wrote a little C program that does any arbitrary number of iterations of the rest of the season and examines the results to determine the championship game participants. The outcome of each game is determined randomly — a random number is selected from 0-1 and used as an index against the a priori probability (as made up by me) that each team would win the game. I.E., if you believe Northwestern has a 70% chance of beating Minnesota, then any random number from 0 to .7 implies a Northwestern win. I used the c rand function with a time-based seeding on each run, please no complaints about the quality of my random numbers, I am just simulating football games for gosh sakes. I toyed around with the a priori probabilities a little to see how sensitive the outcomes were. And then at the end of the season, I apply all the tiebreakers if necessary to see what teams represent the divisions in the title game. I ran the simulation 1000 times, and a few runs of 10,000 trials just for yuks. The 1000 run simulation

So — the Legends division. Not surprisingly, due to the weakness of their remaining schedule, MSU is the title game rep ~2/3rds of the time. Nebraska about ~1/3, and Michigan picks up a smattering (1%) of appearances. Since Michigan has already lost to MSU, if they ever end up tied in the standings, MSU always takes the spot. Michigan needs MSU and Nebraska to both falter (and can control the Nebraska since they have yet to play) and also needs to beat OSU, Illinois, Iowa. A tough road. This weekend’s play won’t shake things up much, the most interesting game is the Michigan-Iowa game, I would have picked Iowa a week ago but losing to Minnesota has shaken my faith in the F(erentz) Troop. The Nov 12th weekend will be more entertaining — MSU@Iowa, Michigan@Illinois, Nebraska@PSU. And then Nebraska@Michigan Nov 19th. It seems unlikely that Michigan will still be in the hunt by the time of the Ohio State game.

The Leaders division is far more interesting. PSU has the inside track, no surprise being up 2 games on everyone else at this point. They take the title game spot 30-60% of the time, depending on how you view the likelihood of them winning their remaining games. If you think they are a slight favorite in all their remaining games, then 60%. If you think they are a modest underdog, then 30%. OSU has a surprisingly good shot, 20-25%, depending on how you rate their odds against Michigan and PSU, and because they are in a good tiebreaker position having beaten Wisconsin. Wisconsin picks up the pieces and has the weakest chance due to the OSU loss. The Nov 5th weekend will likely teach us nothing as PSU has a bye, OSU has Indiana, Wisconsin has Purdue. The Nov 12th Nebraska@PSU game is one to watch, and then Nov 19th with Wisconsin@Illinois and PSU@OSU is a defining weekend. There is a very real chance that on the final weekend, all 3 teams need a win to make it to the game — OSU is at Michigan and PSU is at Wisconsin, so that should be a great day.

I’m struggling to understand why I would ever use iCloud storage. After a couple days of tinkering, I have two sets of data in iCloud — device backups, and Pages/Keynote docs.

• I really don’t get the value of device backups. My apps are all recoverable from the iTunes store. I use primarily apps like Evernote that already store their data in the cloud so there is minimal non-replicated data on my iPhone and iPad. Music isn’t backed up, I will need iTunes Match for that some day. My photos aren’t backed up in iCloud, that is not something that is offered at all (and besides the photos on my device are a fraction of my photo content, I use smugmug and other paid services to back up all my photos). So what exactly is in these device backups that iCloud stores? and why is this substantially better than backups stored on my Mac — when will I ever use these backups? In sum — I’ve been explicit about choosing apps and configuring apps so that all my valuable data and state info is replicated and in the cloud, so that I don’t care if I lose a device (and can use multiple devices). So why should I care about device backups?

• The other files in my iCloud storage are docs. I have Pages and Keynote docs in iCloud from my iPad. If I was purely a Mac person, and didn’t collaborate at all with people in my office and business partners who use Office, then maybe I could just use Pages/Keynote on the Mac, and the iCloud doc storage might seem pretty simple. But I use a PC sometimes to edit my docs. And so I use Office so that I can work on my Mac or PC. And so that I can, with no fidelity loss, work with my colleagues on docs they have created in Office. I guess I could still move these docs in and out of iCloud storage, but if I am going to go to the trouble of moving docs around, why don’t I just move them into box.net or dropbox? They both have great iPad and iPhone interfaces, they work with Pages/Keynote on the iPad, I get 50G free on box.net, they both offer sharing options, I can create folders in them to organize my docs and control my sharing (Seriously, iCloud, no folders??), they let me store any kind of doc, they have great Mac/PC clients so that I can sync my collection with local folders easily, etc etc. If iCloud didn’t have the Apple brand, we would all be laughing at it.

• iCloud claims to store your music but practically doesn’t. I have 16,000 songs, 88G of music, in my iTunes library (and flac versions of all this but not in iTunes). 99% of it is from ripped CDs or purchased in mp3 format outside of iTunes. None of which iCloud handles, I have to wait for iTunes Match.

• iCloud stores your photostream but I’ve already talked about why this isn’t very useful to me.

• I don’t care about mail/calendar/contact backup as all mine is already stored on my Exchange server or Gmail server.

So iCloud storage is substantially worse than leading competitive alternatives for document storage; its only unique benefit is device backup, which I can’t figure out why I’d use; and it’s other features don’t really solve any problems. I am sure Apple will improve iCloud over time but as a storage solution it is underwhelming. Am I missing something? Does anyone find iCloud storage to be hugely helpful?

OK like the rest of the working world I spent hours yesterday trying to upgrade my iPhone4 and iPad2 to iOS5. About a dozen retries for the phone, maybe half that for the iPad, and I finally got there. Not a great experience but no harm done, just a half day of my life wasted that I will never get back, Apple.

So now what? Well my iPhone 4 seems a little zippier but I suspect that is largely due to grinding the old OS off and laying down a bright new clean install. I like the tabs in the Safari. The Newstand seems like an utter waste and sadly cannot be off hidden in an “Utter Waste” folder, thanks Apple. Notifications are cleaner. Renaming the iPod app to Music is good.

and iCloud? Well this is just confusing. Settings spewed all over the control panel — in the iCloud section, but also in the mail/contacts/calendar section, the photos section, the notes section, the store section. Much discussion online about how to make this all work with exchange and how it does or doesn’t work with outlook — for instance http://daggle.com/outlook-icloud-google-calendar-sync-2748. I’ve no idea where things are actually stored in the cloud — the photostream for instance that I have turned on, where is it, can I go see it at a URL? Or Notes — they are associated with an account now, my gmail account. So when I create a new note does it go somewhere in the cloud? Where? The only thing that my cloud control panel lists as being stored is a backup of my phone — why exactly do I want to do this, I never had this in the cloud before, why do I want it in the cloud now?

The design compromises of iCloud — storage limits, and trying to work with a bunch of existing cloud services — seem to have led to a really fractured, incomplete experience. Not all my stuff is in the cloud, what is in the cloud is spewed across many services, and I don’t really know where anything is. Yay.

UPDATE: Ok, new Notes show up in a gmail folder named Notes. Which seems strange, why would I want my notes there? And not in Google Docs or Dropbox or Evernote or … ?

I’m playing around with ifttt.com and it is intriguing. I remember an earlier effort, yubnub, that I always found compelling. A general script interface to all my Internet data and services so that I can do interesting things across sites seems good. 

I remember the evolution of Lotus Notes. A super general collab platform that let you store anything, write nice scripts and forms on top of it. The generalness of the platform was appealing and a certain set of early adopters went for it. But there were many more customers who didn’t want to create their own collaboration apps, but needed some pre-built apps. And so then Lotus made the “nifty fifty” most popular apps available — email, calendaring, candidate tracking, simple CRM, etc etc etc. And that was good, and more customers bought it. But ultimately Notes got washed out of the market by Microsoft Exchange, for many many reasons. But one simple view is that, while Exchange was a collab platform too (although terrible to code against), Exchange really focused on the high volume apps of mail and scheduling and just made those apps work. And that is all most people really needed.

Competitively the ifttt.com guys need to be very cognizant of cherry picking. While it is great they have hundreds or thousands of canned scripts, I don’t need hundreds, I only need a couple. And that is probably true for most users. And if the couple that people need a common across large groups of users, then some competitor can sweep in and just do those couple scenarios really really well and ifttt.com will remain a niche tool. I’d bet that they will have to build a lot more code on top of their platform to make sure that the top scenarios are really slick and easy to use, to avoid losing users to alternatives.

For instance I can already pretty easily use a wordpress plugin to MIRV content over to twitter and then to facebook. Will I flip over to ifttt.com for this or will I keep using the solution that someone has polished and made fit into wordpress? I suspect I’ll use the one that fits really well in wordpress. Now if the ifttt.com guys wrote the code to provide an ifttt.com plugin for wordpress, that would be interesting…

This year I’ve had a chance to experiment with tools for compute intensive applications. In particular, tools that harness the profusion of inexpensive CPU/GPU cycles available — OpenMP for multi-threading on single machines so that multiple cores can be leveraged; MPI to distribute compute load over clusters of machines; OpenCL for handing general purpose computation off to a graphics processor. And then on top of these tools, NumPy and SciPy for scripting and visualization from Python. The amount of excellent computational software which is now available is amazing, these capabilities would have cost immeasurable amounts of money just a decade ago. And the first time I tied together a cluster of machines or yoked up a GPU and did a massive computation, and then displayed the animated results using Python — what a great feeling! The ability to attack really hard, really large problems is better than it is has ever been.

But what a nightmare of housekeeping. Breaking up computation into threads and spreading it across multiple cores with shared memory and file system is tedious and error-prone — hand-offs between threads create opportunities for many errors. The work to break up and manage the computation load across multiple machines is even more mind-numbing and error-prone, and now the lack of shared memory and files are additional complications. Using graphics processors is even more obtuse, with their funky fractured memory spaces and architectures and limited language support. And getting all the software piece parts running in the first place takes a long time to work through all the dependencies, mixing and matching distributions and libraries and tools, and then getting it all right on multiple machines. And then you get to maintain all this as new versions of libs and runtimes are released..

But again the results can be stunning — just look around the web at what people are doing in engineering (Youtube video), life sciences (Science Mag article), or any of a dozen other areas. Harnessing multiple cheap processors to perform complicated modeling or visualization can have huge payoff in financial services, bioinformatics, engineering analysis, climate modeling, actuarial analysis, targeting analysis, and so many other areas. 

However, it is just too darn hard to wield all these tools. The space is crying out for a cloud solution. I want someone else to figure out all the dependencies and library requirements and spin up the correctly configured virtual machines with all the necessary componentry. And keep that up to date as new libraries and components are developed. I want someone else to figure out the clustering and let me elastically spin up 1, 10, 100 machines as I need to, and manage all the housekeeping between these machines. I want someone else to buy all the machines and run them, and let me share them with other users, because my use is very episodic, and I don’t want to pay for 100 or 1000 or 10000 machines all the time, when I only need the machines for a week here and there. Maybe I want to run all my code in the cloud, or maybe I want to have all the VMs and clustering info delivered to my data center, but I want someone else to solve the housekeeping and configuration issues, and let me get to work on my problems.

Amazon is doing some great work in AWS with their HPC support (AWS HPC support).
Microsoft has made a commitment to provide scientific computing resources in the cloud (NYT article). There is a lot of great academic work happening (ScienceCloud2011). But the opportunity is out there to do a lot more.

• Soluto. Soluto seems right up my alley — focused on simple common frustrations that we all have, promises to save me time. Install is a breeze and I really like the super sparse interface — such a difference from the overcomplicated software from Norton, etc. The software feels very light and the interface reinforces the promise of simplicity. And it does seem to make tuning up boot time simple, it pretty accurately understood all my boot processes and gave me reasonable suggestions. It’s browser diagnosis was less helpful, it found very little in Chrome to improve, and it didn’t look at my firefox install at all. Not sure why. Anyway, worth a trial and I will be interested to see how they go. A challenge they will have is getting users to pay — solving my boot speed issues is nice, but I only need to do that once and I have no enduring reason to keep on running the software — and thus am not going to pay much for it. They need to figure out a way to deliver me value every day. The only guys in the utility space that have done this are the virus/malware protection guys, who have latched onto consumer fear (and probably stoke that fear).

• Splunk Python interface. Really curious to play with this (disclosure, Ignition is an investor in splunk). I hadn’t installed splunk in a while, installs super simply on Mac and Win. and wow what a firehose of info you get from Splunk about your system. Next up, tie to python and try to write some simple scripts. A lot to play with here.

• Calculize. Kind of like Matlab, in the browser. Might be useful. at times.

• Keymando. Love the idea of hotkey utilities but I always seem to drift away from them. Because I can never keep them in sync across all my machines. And so I will probably never install this. But noted here in case I try.

• timely.is. Rand likes it which is a good sign. If I cared about readership and impact of my tweets I think I would certainly give this a whirl.

• Onswipe. This seemed really cool, but I thought it was basically a wordpress theme. It isn’t tho, it grabs your wordpress data and puts it behidn a new url. and it seems to be dependent on categories which I don’t use. so I will wait.

UPDATE: nice simple tip from the Keymando guys — use Git or Dropbox to keep Keymando settings in sync across multiple machines. This is a simple obvious brilliant thing I should do in general for my work and home Macs.

• How to unhide your Library in Lion — chflags nohidden ~/Library. Yay.
• Making desktop web apps with Automator. All kinds of goodness in here.
• BBEdit 10 is out. Purchased.
• Marked seems like a super useful companion to BBEdit.
• Billguard is now free-er.

I’ve wished for years that someone would come up with a software-defined lens. A surface that would capture all inbound photons and let me decide later about focus, depth of field, etc. 

It looks like Lytro has done it or something on the way towards it. Hope it is reality! Put my name down for one.

Been spending a lot of time playing around with GPU programming for scientific computing the last couple weeks. Fascinating stuff, GPUs are computational beasts. Some observations:

• If you want to get into it, GPGPU.org has boatloads of great info — news, tools, definitions, primers, etc etc etc. The place to start.
• There is a good chance you’ll end up using OpenCL as the device- and platform-independent interface to GPUs. Khronos.org has tons of great info and in particular, the OpenCL Reference Card. Good stuff.
• The OSX platform has awesome support for OpenCL within Xcode. Very easy to get up and going. Great sample code up at the Apple Developer web site.
• Also tons of samples from Nvidia.
• However…you may quickly hit a dead end on OSX because only the most expensive Mac Pros come with GPUs which will support double precision, and double precision is kind of necessary for scientific computing. Info on which Nvidia processors support double precision here. I could go whack around and build my own double precision math libraries for unsupported GPUs but what a pain that would be.
• So onto a PC, I happen to have one with an ATI HD 57xx which will support double precision. WAY harder to get working OpenCL code working on a Windows PC tho. After much wandering around, the “AMD SDK“http://developer.amd.com/gpu/AMDAPPSDK/Pages/default.aspx seems to be the best way to get working buildable OpenCL sample code. The most freaking obtuse make files ever tho, I am ripping them apart. But if you start with one of the sample code bases and duplicate it for your use, it works. (C++ by the way).
• However now I am currently blocked by limitations in the trig function implementations. Some discussion online that suggests that they are single precision only. And even the single precision results seem to have crappy precision. I will definitely have to build my own.

UPDATE: a friend points out that Amazon also offers an EC2-based instance with GPU capabilities. Worth a look

• iPhoneTracker. Cool toy to see where you’ve been and feed your feelings of paranoia.
• Portal 2. Of course.
• Qwiki. I was kind of excited about this, but I can’t make my own Qwikis? Excitement way down.
• Acorn. Haven’t bit yet but I’d love something less obtuse than Photoshop.
• Blogsy. Seems like a brilliant WordPress front end.
• Issue Bucket. Nice little frontend to bitbucket.
• You Gotta See This. Stupid little stylized panorama camera app for the iphone. Fun.
• Office365 beta. The individual apps (word, xl, ppt) are nice and well done. The portal gluing them all together with email and calendar is strange and confusing — two URLs, yet another ID different than my existing ID used at all msft sites, an insistence on downloading software. Chalk it up to beta.

Saving for later reference….

• Netlib
• NIST
• Trilinos
• PETSc
• OpenCL
• Nividia OpenCL
• Apple OpenCL
• NERSC. ACTS stuff dead?

• gfxcardstatus to let me fiddle with macbook pro graphics hardware. which is proving to be problematic. Why does the browser (Chrome) require the high end power-consumptive nvidia chip? Seems like this feature of the macbook is a waste if the browser is always going to force the power hungry chip on. OK hmm, this might be just a Chrome issue as Safari is staying on the intel chip. gfxcardstatus is great for examining status and dependencies!
• techdygest. Might be a little too digested. But worth a try.
• daytum. I want to love this but too hard to get going. There needs to be some easier way to populate it with personal data. 
• socialeyes and dailybooth. There is something intriguing about the front-facing camera. I suspect there will be a lot more software written around. What will be the first front-facing camera game? (Ignition is an investor)
• greplin chrome extension. Search of my content seems super fast, i am intrigued. (Ignition is an investor)

• Printopia to enable printing to any printer from iphone/ipad (via Tidbits(TidBITS:+Mac+News+for+the+Rest+of+Us))
• Textastic code editor for the iPad (via Read/Write Web(ReadWriteWeb))
• Postbox email alternative for OSX.
• Momento, a diary app. (via TechCrunch(TechCrunch))
• A whole bag of iphone apps for engineers.
• Calvetica replacment for iphone calendar app.
• boxcar. Don’t really grok this one but people seem to like it.
• Firesheep(TechCrunch)
• Bit.ly bundles(GigaOM:+Tech)
• mynameise
• GoMiles
• One man’s view on essential programmer utilities
• Google Public Data Explorer

• Poor Halo play prompts stabbing threat. Doesn’t seem unreasonable. I’ve heard campers threatened with worse.
• How Secure Is My Password. No idea how accurate, but fun. 17 thousand years for my typical password.
• Rich on photobooks. I just always use the default in Aperture but perhaps I should branch out.
• AR.Drone. Why don’t I have one of these yet.
• Declining energy quality as recession cause. An interesting way to look at things. Not sure it actually makes sense tho.
• Now you can swap useless Amex reward points for useless Zynga crap.(TechCrunch)
• Snoopy themed Windows tablet. Take that, Apple.
• Spiders on Drugs(Design+Milk). We are asking for some serious payback from spider nation some day.
• Umpteenth article on the death of cable TV(TechCrunch), yawn. Until I can watch HD live sports without stuttering I am captive to cable/dish. Going to be a while.
• Charles on breaking up MSFT. Good as always.
• Habitable planet found?
• Languages you’ve never heard of(Gadgetopia). In the future, we will all have our very own programming language.
• Topologist suggests new form of matter. For most of our history we’ve used the forms of matter that nature gave us. It is interesting to observe and think about what we can create as we gain mastery over atomic organization.
• One man’s indictment of iTunes. The thing is a giant hairball of software.
• Exercise and aging. Crap I need to get after it.
• Show Me What’s Wrong. Super useful.
• User experience of F1 telemetry. Always impressed with the amount of money spent on racing.
• MacPaint and MacDraw source code. Nostalgia.

Kicking myself totally on this one, VMWare has been on a tear. Paul is a great guy, he has been hiring great guys (who wouldn’t want to work with Paul?), they’ve been acquiring lots of interesting assets.

And fundamentally they are on the right side of history. Paul has always been insightful and articulate on strategy and he says it well in this techcrunch piece : “The innovation in how hardware is coordinated today and the innovation in how services are provided to applications is no longer happening inside the operating system.”

This is dead on. You can debate whether VMWare will be the primary beneficiary of this trend versus other cloud providers, but the shift is undeniable.

• Firefox is feeling increasingly bloated, maybe because I’ve got a bunch of plugins jammed in. But trying out Camino(TidBITS%3A+Mac+News+for+the+Rest+of+Us)&utm_content=Google+Reader on the Mac, seems cleaner and lighter.
• Shuffler.FM. Eh, streaming music just doesn’t work for me. My primary listening time is while driving and I need music that I can put on an ipod or cd. When I am at an actual computer I am too busy doing other things. But I like music discovery tools and guides, I just don’t want them bound into streaming.
• GIT for the lazy. Perfect for me.
• Terminal tips and tricks for OSX and in general SuperUser seems helpful.
• I want to love WiseStamp but I don’t get email addins that assume you are only sending email from a browser. iPhone? iPad? OSX Mail? How can I commit to this thing if I can’t use it consistently? Sigh.
• I’m super late to Windows Live Sync but it is very useful. I do have a quibble with the naming, once upon a time MSFT was confident enough in its products to give them simple iconic names — Word, Excel, Windows. The company seems to have lost its confidence in products and jams these crazy names on them to try to ride on the coattails of other products. Mistake.