ReachTech Blog

Just came back from the Embedded show in San Jose where I talked to a salesperson from Densitron (www.densitron.com). He told me about a pretty neat development: resistive touch panels that have an ultra-thin glass as the top sheet instead of the usual plastic. The glass is so thin that it can be flexed by a finger or pen just like the top sheet of a standard resistive touchscreen. Because it is glass, it is more durable and cleanable. It is ideal for medical applications where traditional capacitive touch doesn’t work well with gloves. It is new and likely expensive, but looks like a good option to have available.

Everyone wants a reliable system, right? Well, in my humble opinion, the biggest cause of unreliability is dynamic memory allocation. Ever have Windows freeze up? If you are lucky it may tell you it has “insufficient resources to complete task” which means there was an allocation problem. If a program that allocates memory on the fly cannot do so for any reason, the program has no option but to stop running, either gracefully or more often with a thud.

For programmers, dynamic allocation is a great way to handle the unknown. How many users will the system need to support? Don’t know? Then allocate memory as needed. What’s the maximum number of lines in a file? Allocate again. The result is a system that cannot be thoroughly tested since with each additional user, open file, added link and so on, the combinations increase exponentially. To make it worse, the amount of work required to gracefully shut down an application when a low level allocation fails is tremendous, so most applications just abort.

The other problem with dynamic allocation is that as users are added or as more files are opened, the system often has to slow down every now and then to either de-fragment the memory pool, or wade through the dynamic links to do the requested task. The result is that the system response time varies. And the variation doesn’t depend on how it is currently being used, but how it was used in the past. What’s the difference between a calculator and a computer? The calculator will always take the same amount of time to deliver a result, and will demonstrate none of the pauses, or “stuttering” that computer users have accepted as a way of life. When something works in the same way each time it is used, it is inherently “looks” more reliable.

In case you were wondering, no, I don’t prefer DOS to Windows XP, but DOS was a heck of a lot more reliable….

Selecting a small LCD panel for an embedded application? Here are some tips.

1. For technical information, the manufacturers and their sales representatives are generally a better place to start than the traditional broad-line panel distributors. They have more current information and are generally more focused. Go to the panel maker’s web pages (e.g. americas.kyocera.com, necelam.com, sharpsma.com) and look for “Sales Representative”. The panel distributors have to ask these folks for the same information, so you might as well go to the source. Ask for the current product roadmap; this may tell you whether that panel you’re looking at will be around for the next few years. In general, vendor websites do not have the most up-to-date information.

2. For pricing, ask the manufacturer’s representative for the “recommended resale price assuming registration”. This is somewhat like the “dealer invoice” price for cars, and will give you an idea of what you should be paying for the panel when bought though distribution. Ask them which distributor sells the most of the specific panel you want, and which ones are the easist to work with. If you want the best price, call all the distributors and use the recommended resale as a starting point for negotiation. If you want best availability, go for the distributor that sells the most; generally they will give you a good price as well.

Future topics:

- why panels go obsolete and why they all have different connectors in different places….

- why panel specifications don’t tell the whole story

Welcome. Why blog? Well, it offers an informal way to discuss things that don’t fit neatly into product documentation or FAQs. The blog is written by various members of the Reach team. We talk to small LCD panel users every day, and some of the conversations’ topics will show up here.