My Thought Bin

There are 4 options to consider in developing a new device: An integrated device that supports the core functionality needed for your application. Advantages - no certification costs, proven performance, shortest time to market Disdavantages - higher per unit cost & limited flexibility in the functionality Examples - Typcial Deployment Range - 1 to 1000 Vendors - depends on application An end product is a standalone wireless product that can be integrated with other hardware to form an integrated device. Advantages - no certification costs, proven performance, short time to market Disdavantages - per unit cost Typcial Deployment Range - 1k to 10K Vendor - MultiTech’s Socket Modem www.multitech.com An integration with a module Advantages - lower per unit cost, flexibility in form factor, reduced operator certification (assuming the module is pre-certified by the operator) Disadvantages - certification costs (PTCRB - $20K-40K), operator certification may be required (typically no cost) Typical Deplotyment Volume - >10K Vendors - Cinterion, Telit, Sierra Wireless, Ericsson, Motorola, Enfora, Simcom An integration with a chipset Advantages - lowest per unit cost & flexibility in form factor Disadvantages - certification costs (PTCRB - $60-80k, Operator - up to $500k), long time to market 12-24 months & high level of wireless expertise is required for development Typical Deployment Volume - 1 million Vendors - Infineon, Qualcomm, Marvel, MedioTek, Texas Instruments, ST Ericsson, Freescale As you can see, it really depends on where you are with your business model, but there are options to meet every need. Also, I highly recommend contacting your operator to find out their requirements.

Well, as with any technology someone owns the IP and you’ll need a license to develop with it. Most of the time its Qualcomm and you need to get a IP license to the technology you are trying to implement with (i.e. WCDMA) and the IP license usually requires a substantial up front cash investment (think millions) and/or a minimum order quantity. Honestly, unless you are close to or in “cell phone volumes” don’t even bother because by the time you subsidize the cost of the IP licensing, NRE for development, minimum orders, and everything else unless you are truly shipping 7 figure volumes you’ll end up losing money. Our firm has been in this space for a decade. 90% of our new customers always want to start with the chipset level and 99.5% of them change their mind once they handle the due diligence. I’ll throw some ball park numbers at you to show you why: IP Licenses: $1-$3 Million RF Certifications (Full FCC, PTCRB, then carrier certification): Approximately $1.3 Million Hardware Design NRE & Prototyping Costs: $300K-$1M depending on what you’re trying to build. Out of the hundreds of projects we’ve worked on in the past 10 years, only two have ever truly gone to an integrated chipset (base band design.) Really, by going with an integrated chipset you are doing exactly what Cinterion, Telit, Wavecom, and the rest of them have done. You are literally creating your own module and that gets really expensive really quick. Also your time to market is a lot slower as all the certifications take a lot of time. In order to go with an integrated chipset you need to have some very experienced engineers on staff or be willing to pay the NRE costs to sub out the work. Another issue is what type of integrated chipset are you looking at and is your run rate there to keep up with the end of life cycles because handset chipsets go EOL a lot quicker than other silicon. All in all, unless you are ready to spend about $5 Million and the next 10 months dealing with headaches I’d stay away from an integrated chipset.

Process should be written by those who are not only intimately experiencing the pain of a lack of process, but who are also experts in the culture.

C:\mklink /D C:\TestFolder C:\Users\Geek\TestFolder
symbolic link created for C:\TestFolder «===» C:\Users\Geek\TestFolde

useful for calculating seconds since epoch

This is a classic Fitts’s Law problem, and Fitts’s Law states that the ease of hitting a target is a function of both the size of the target and the distance to get there. So you want bigger targets on bigger displays, because there’s more distance for your pointer (in this case, your finger) to travel, but you can get away with smaller targets on a smaller display because there’s less distance to travel.

To OEMs, it seems creating a family of products means nothing more than simply making a set of devices look similar. Hogwash. The reason Apple has been so successful is because its family of products operate better as a unit, with cross-platform features (iTunes, iCloud, iMessage, and so forth) that no competitor can rival. The experiences need to be designed together.

The web kills off other protocols because it has something most protocols lack: a simple way of labeling every available item. Every resource on the Web has at least one URI. You can stick a URI on a billboard. People can see that billboard, type that URI into their web browsers, and go right to the resource you wanted to show them. It may seem strange, but this everyday interaction was impossible before URIs were invented.