Tuesday, October 30, 2012

Kickstarter, Parallella And 21st Century Brand Building

Victorinox Swiss Army knife, photo taken in Sw...
Victorinox Swiss Army knife, photo taken in Sweden. This is a Mountaineer model. (Photo credit: Wikipedia)
I'm a technology junkie. Its a bug that bit me when I was still in my teens and managed to buy my first computer with money I'd earned from a summer job.

Business is a passion that developed a little later in my life when I started wondering why some products succeeded and others failed. In particular I was confused when technologically superior products somehow lost out. This made no sense to me at the time and like most technically inclined people it drove me nuts.

It turns out there are some good logical reasons why companies and products fail. For instance not knowing what problem they are solving or spending their money in ways that don't actually help build their business. But there are bad reasons as well. One of the bad reasons is not being able to get the funding needed to take an idea to the finish line. This is particularly true when the idea in question isn't bad, but rather different enough from what has come before that the people with the money have trouble wrapping their heads around it.

Kickstarter has been around for awhile now but I'd never quite gotten around to checking it out. It turns out they provide a solution to the problems mentioned above. Kickstarter does this by providing a marketplace for ideas. If your  idea catches the attention of enough people you get the money and sales you need to take that idea to market.

I have a background in HPC (High Performance Computing) which is why the recent Apapteva Parallella Kickstarter campaign caught my eye. Adapteva designs and manufacturers parallel processor architectures. From what I can tell they are a very small company operating in a marketplace that is both competitive and cost conscious. Super computing is sexy, but very few companies make money at it and the margins are thin when they do.

Adapteva's architecture should eventually scale to 4096 cores on a single piece of silicon. That's a lot of processing power but there are issues. The current offering is only sixteen cores and each of those cores has just thirty two kilobytes of memory. I'm primarily a systems guy so I don't have a deep understanding of HPC programming but I do know that having so little memory significantly constrains the kinds of problems this iteration of the product will be capable of solving. Adapteva apparently understands this as well and chose to position their Parallella board primarily as an educational tool.

The Parallella board couple's Adapteva's sixteen core parallel processing chip with a SOC that includes two 1GHz ARM A9 processors an FPGA and other useful things like wired networking and USB 2.0. The ARM processors run Linux and the FPGA is used to interface between the Adapteva parallel processor and the outside world.

If you're not a technically inclined person just imagine this as a kind of Swiss Army knife of small computing platforms.

The Adapteva board will be credit card sized and cost just under $100. That is a killer combination.

Kickstarter isn't only about money though. The Parallella campaign was in trouble most of the way as it was not seeing the kind of growth in funding it was going to need to meet its goal of $750k. If you don't meet your funding goal you get nothing so this was a serious problem. What they were able to do, particularly over the final week of the campaign was to slightly refocus their target audience and release additional information that energized and excited their existing backers while encouraging many new backers to buy into the campaign. When the campaign ended they had raised nearly $900k

This ability to get quick feedback and adjust the message is one of the advantages of a Kickstarter campaign. That kind of market research is reasonably easy for web based businesses as the Lean Startup methodology espouses/shows but finding something equivalent for a physical product is much tougher. Kickstarter fills the niche nicely.

Kickstarter also lives up to its name by enabling people with products to build an initial customer base. Kickstarter projects such as Parallella are raising money in large part by selling the promise of a product being delivered to the backer in the future. Not all backers buy in at this level but it is an option for both backers and projects. Adapteva used this to pre sell thousands of their Parallella boards. All this helps build the company and its brand.

Funding, market research, customer acquisition and brand building can all be accomplished by a well designed Kickstarter campaign. Kickstarter isn't a non profit so all this useful stuff does come at a cost. 5% of the proceeds go to Kickstarter with another 3-5% going to the payment processors. This seems like a very reasonable cost to me. In addition project owners retain all rights to their intellectual property.

I can't really think of any downsides from the project side. There is a potential downside if you are a backer in that a project creator may never deliver on their promises. Investing is always a risk though and I suspect we'll see laws put into effect over the next few years to better cover how that situation could be handled.

I decided to back the Parallella campaign because I found the idea intriguing and I wanted to get my hands on one of their boards. If things go well I should have one by next May.

Related Articles:
Enhanced by Zemanta

Wednesday, October 24, 2012

Apple Products Not Over Priced (But the iPad mini may be)

Steve & Apple Inc.
Steve & Apple Inc. (Photo credit: marcopako )
I have an auto translation mechanism in my brain that changes "X is too expensive" into "X costs more than I'm willing/able to spend". I'm not being harsh here either. We're all a bit narcissistic so the tendency to project our own feelings onto the rest of the world isn't surprising.

In the tech world Apple is the company that gets the most flack for their pricing. This has been going on for decades now and will almost certainly continue for as long as Apple is a viable company. What most people fail to understand is that no company or product can realistically appeal to everyone. Apple is no exception. 

I could write a bunch of text here about Apple being a premium brand and how that impacts their ability to price but I've said it all before and its really not my primary point right now. Suffice it to say that the billions of dollars Apple has in the bank server as excellent evidence that their products are as a general rule not over priced. With one exception I think they are dead on again with their latest slate of new products.

To be blunt, the iPad mini looks like something Steve Jobs would have killed circa mid 2010 when he was still very much involved in the day to day operation of the company. It's woefully short on the technical side which isn't fatal but what is worse is that at the prices Apple is charging it is short on value. That won't stop people from buying it but over the mid to long term it is much more likely that people's good feeling for the company are going to be eroded.

The problem Apple appears to be addressing with the iPad mini is that competitors like Amazon and Google are selling a lot of units in the smaller form factor tablet market. Smaller tablets are much easier to carry around and if people have them they'll be more likely to partake of whatever ecosystem their tablet supports. Apple is primarily a hardware company but they make a fair chunk of money off their apps and media sales. Seeding the field to their competitors was not something they could afford to do. 

Lets assume Jobs was steadfastly opposed to the smaller form factor tablet. The man was a genius but he was also been known to be stubborn on things that he probably shouldn't have been. Shipping the original NeXT computer with a very immature optical disk technology as its only storage being one of many examples I could cite. If this is true it would mean Apple has had just over a year to get the iPad mini to market. That isn't a lot of time and they would have had to take shortcuts. Like perhaps reviving a design that Jobs had shot down. That would explain the sadly out of date tech specs. This thing is essentially a smaller more refined version of the iPad 2, a product that is now two generations behind Apple's current full sized iPad.

So maybe this is a stop gap and Apple didn't want to release it at a price point they would have to maintain for future generations? If this theory is correct we could see a next generations iPad mini in less than a years time that supports a retina like display, some variant of the A6 processor and at least a gig of memory. A device with those kinds of specs would be justified to sell for the price points Apple is currently asking.

Personally I don't see any reason to purchase this initial version of the iPad mini. Apple underwhelms badly on the technical side and are asking for an even bigger premium than normal on the price side. The math just doesn't add up.

They'll likely still move a lot of units but I suspect customer satisfaction on this first generation version is going to be lower than is typically the case and that could start to hurt them in the longer term. A company as successful as Apple can afford an occasional miss but they can't make a habit of it. This is particularly true now that Jobs is gone.
Enhanced by Zemanta

Monday, October 15, 2012

NVidia And The FOSS Community

English: NVIDIA Tegra T2 based Embedded Comput...
English: NVIDIA Tegra T2 based Embedded Computer Module in SoDIMM format by Toradex, Switzerland (Photo credit: Wikipedia)

Background

I've observed over the years that the FOSS (Free/Open Source Software) community loves to hate NVidia. Linus Torvalds, creator of the Linux kernel famously gave NVidia the finger and dropped the F bomb when describing how he felt about the company earlier this year. Torvalds isn't known for mincing words but at times he's also been less than popular with the hard core FOSS crowd for failing to be completely pure in his adherence to Open Source ideals so its probably safe to describe that event as a low point.

My roots in the FOSS community go back a long way. I did some work on porting the shadow password suite to SunOS at one point and wrote or edited the bulk of the early MySQL documentation. I also did a lot of work on a now defunct open source task tracking system called "request". I love the ideals of open source software but there are times when the sense of entitlement expressed by some adherents annoys me. Companies and individuals have to generate enough revenues to get by. If they don't bad things happen. While pictures of people with signs saying "Will code for food" are amusing the reality of being in that situation isn't.

Balancing the interests of FOSS ideals with the need to provide for ones shareholders or family is tricky, complicated and absolutely necessary.

In researching this article I was pleasantly surprised to see that there have been changes recently that may signal an improvement in the relationship between NVidia and the FOSS community. This will hopefully be a good thing for both sides.


The Main Issue

The central beef of the FOSS community with NVidia has centered on the companies refusal to release an open source driver for their video cards. Making matters worse NVidia has also refused to release technical details that would enable others to do so; that is starting to change though. NVidia recently released technical details on the 2D portion of their Tegra architecture. Some sources claim there will be a release of information on the 3D portion of the Tegra line as well. I couldn't find anything on NVidia's site at this time but I did find the following text at

http://developer.nvidia.com/mobile/tegra-documentation

The Tegra Technical Reference Manuals are technical documents targeted at those working on open source or other low level software projects that use or target the Tegra 2 or 3 processors.

Unfortunately at this time, NVIDIA cannot provide access to the TRM (or accompanying Embedded Controller Interface documentation) but please check back to http://developer.nvidia.com in the near future for more information


The potential implication here are very promising.

The question a lot of FOSS adherents struggle with is why doesn't NVidia just open up their source and documentation right away? There are probably a lot of reasons for that but if I were to guess, the primary reason probably has to do with being a publicly traded company with shareholders and employees that expect to be rewarded for their investment of money and/or time.

NVidia likely has a number of trade secrets as well as partner code incorporated into their hardware and software. Trade secrets aren't secrets if they are documented in text or code that is widely available. In an ideal world there would be no need for such secrets but the world we live in is far from ideal.

If I'm correct, releasing code or documentation would require a non trivial amount of work on NVidia's part to sanitize and remove anything they either didn't own or didn't want to reveal. That work would need to be done by valuable staff that could be working on other projects which would create both an up front cost and what is referred to int he business world as an "opportunity cost". Companies incur opportunity costs when they choose to work on one thing rather than another. Smart companies make the right choice on where to focus their resources more often than not. There has to be a strong business case for NVidia to justify making this effort. If there isn't, it simply won't happen.

Yes, other companies have gone down this path successfully but that doesn't mean NVidia must or should. Another consideration is that every decision a company makes is going to reflect on its brand. NVidia has in the past essentially chosen to annoy the FOSS community rather than lose some control of how their product is perceived from a performance and reliability perspective. Is that a good choice? It depends on who you are I guess. On balance NVidia has apparently felt it is.

I covered this in passing above but want to reinforce the point. Time to market in the technology sector is critical. Being first at a particular performance point can be a huge advantage. AMD was able to put Intel back on its heels for a few years by releasing processors that provided a 64 bit IA-32 compatible instruction set and much better performance than anything Intel had. Keeping their drivers closed allows NVidia to license technology from other companies more easily than they could if their driver were FOSS. It also means all driver related coordination can take place within the company. That control likely equates to shorter time to market which translates into money.

Can those difficulties be overcome? Sure but the old missive "If it ain't broke don't fix it" comes to mind. Innovation is good, but change without a strong business case is the road to ruin.


The Larger Picture/The Future

Rumors to the contrary NVidia isn't universally opposed to FOSS. NVCC, NVidia's CUDA compiler is built on the LLVM FOSS compiler infrastructure and earlier this year NVidia released NVCC and contributed it back to the larger community as open source software. That's a big chunk of code with a lot of functionality. Why would NVidia have done that? Only they know for sure but I'm betting they did it because they saw it as being beneficial to their customers and to themselves. Its easy to make a business case when those two things are in alignment.

NVidia has also shown a willingness to express support for causes near and dear to many FOSS community members such as stating they did not support SOPA

If I were to guess I'd say we'll likely see more support for FOSS from NVidia going forward. This is only going to happen though if it makes sense for NVidia. Benefiting the FOSS community at its own expense is not in the best interest of NVidia or the majority of its customers.

Getting something valuable for free is always going to be attractive. Providing something valuable for free is a much harder case to make. At the end of the day that dichotomy explains the bulk of the divide between the FOSS community and companies like NVidia.

Related Articles:
Enhanced by Zemanta

Sunday, October 7, 2012

My $32 Smart Phone

picture of a palm pixi plus
picture of a palm pixi plus (Photo credit: Wikipedia)
Recently my trusty first generation Droid X went missing. The screen has several dead spots from having been dropped one too many times as well so it seemed like a good time for a new phone.

We'd used my upgrade for something else though and I'm not eligible for another until February of next year. Add in the fact that I'm not that excited about any of the current Android phones and buying a long term solution wasn't in the cards. What I needed was something to tide me over for the next three to six months and I didn't want to spend a lot of money.

Amazon is my default search engine when I'm thinking about buying something. Google is a great general purpose search but falls down when it comes to niche cases like sales.

I found some Android phones for under $100 with no contract including the original Droid phone at about $85 but all ran 2.X versions of Android and were likely to be a similar but inferior experience to my Droid X. Oddly the first generation Droid X is still going for $160 or so without a contract. It was a great phone two plus years ago but these days that price seems high.

Enter the Palm Pixi Plus. HP bought Palm a couple of years ago so the phone is an orphan which explains the low price; $32. The Pixi runs WebOS 1.4.5 which is a totally different experience to Android or iOS. WebOS was called PalmOS before HP bought Palm. I just looked on Amazon and its $46 now. Still a good deal and I would have pulled the trigger at that price as well.

I had zero experience with WebOS so this phone also provided an opportunity to learn about a new platform. The version of WebOS on the Pixi is fairly old at this point though so I'm not going to draw any firm conclusions from my experience. HP has a way of screwing things up. It's a company in need of major changes. Having said that, I really like WebOS. It's reasonably intuitive if you give it a chance and easy to interact with.

While the price was a major upside to this phone, there are downsides. The App store is a ghost town and a lot of what is available doesn't actually work once you download it. I couldn't find a free Facebook or Twitter app and I wasn't willing to pay in case what I downloaded didn't work either.

What I do have is access to my email and the ability to text and make phone calls. The screen on the Pixi is small enough that I won't be using the browser much. It also syncs my calendar and contacts from Google which is a nice feature.

Battery life is impressive. In light use the battery meter drops about 1% an hour meaning its good for three days plus between charges if you're just checking your email or texting occasionally. Extended talk time would kill that estimate of course but who talks on their phone these days?

Based on my experience the Pixi Plus was a good good entry level smart phone when it was released. It would have been interesting to see where Palm went from there if they had the cash on hand to stay independent. 

Looking towards the future,  the Samsung Galaxy Note 2 might be my next phone. I like the new split screen capability that allows two applications to run in the foreground at once and I don't think the size would bug me. Another possibility is the next Google Nexus phone. It's nice to have the option

While the Pixi is a closeout it does illustrate an emerging trend. Technology is getting very inexpensive. If you can afford to feed yourself then chances are you can afford some sort of computing device. The real barrier to entry is the cost of Internet connectivity and there are ways around that as well.

There has been a lot of talk about the transition to mobile devices and how they allow us to always be connected. We're also seeing the emergence of ubiquitous computing. There has been talk for a long time of every device in our houses being Internet connected and controllable. I'm not sure that is a good idea but it just might be economically feasible in the next couple of years if it isn't already.

In case anyone is interested in the phone, here's the link. Full disclosure, I'd get a few bucks if you bought via the link.


Verizon Pixi Plus Link
Enhanced by Zemanta