On Mon, 19 May 2008, Steve Nordhauser wrote: > Xiaofan Chen gmail.com> writes: > > > > Microchip rationale, somewhat confirmed by the class attendance, is that > > > it does not make sense to make the effort to support an operating system > > > for a potential incremental audience of just 1% ? > > > TI explained this to me when I tried to buy their DSC chips in low > volume for a commercial project. It was their 80/20 rule. You can make > 80% of the sales with just 20% of the customers. Also, 20% of your > sales can consume 80% of your support resources. Basically, I fell into > the low volume/high support dead zone. They could spend the same > support resources and sell 500K chips/yr to another company or 500 > chips/yr to me. Until Microchip is convinced that some high volume > customers are being lost due to lack of Linux support, there is no real > incentive for them. Overall they are one of the best companies for > supporting low volume and hobbyist design-ins. If all they really > wanted was high volume, they would move their FAEs from writing > application notes to working one-on-one with anyone with an application > of more than 50K chips/yr. Parts would only be available in reels and > evaluation quantities would be evaluated for opportunity before you > could get less than a reel. If I buy 500k chips/yr I would expect a better price per chip than if I only bought 100k and VERY much better than if I only bought 500/year. In other words there is a bigger margin on lower volume. I would assume that there are many more 500/year potential customers than 500k/year potential customers and that the profits for the total 500/year customers could easily justify investing in extra support. I think a more reasonable explanation is capacity. Take the argument to the extream: if they had a surplus of chips they would be chasing every sale no mater how small. They are telling you that you are too small because they don't want to look small. Turning the whole argument around, if microchip became more agressive pushing it's limited chip supply at only the high end, it would be selling on a lower margin. Another point: if you have a product and you need 50k MCUs/year, you're not going to be bothered about linux, windows, BSD, solaris etc. the product is going to be way more important than the tools. Most people would even learn to use a new assembler for that. Regards Sergio -- http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist