On Sat, Aug 27, 2005 at 01:23:47PM +0800, Chen Xiao Fan wrote: > I have no problem with people choosing any programmers. > It is their choice. However I will still consider JDM > type as legacy just like 16F84. It is my observation > and I think lots of the people will agree with me. People > can start with JDM. However it is better to move to > something like Wisp628 later. Adding a external power > supply and a RS232 level converter will probably rescue > a JDM in some cases but then the price will not be so different > from Wisp628. Xiaofan, You keep espousing that argument, but I never see a real justification as to why. Wouter has pointed out that cost effective programmers have their place because they are cost effective. The bootloader approach I discuss pretty much demands an inexpensive programmer. Truthfully even a homebuilt Wisp628 needs a cheap, easy to build programmer. For the sake of future discussion, let's call those types of programmer bootstrap code dumpers (BCD! Funny!) > I do not agree with you that those programmers with PIC > inside have only only one-man support. Here I like open > source implementations like Wisp628. Even Wouter stops > supporting Wisp628 (unlikely now), the source code is > available and alternative software like Xwisp2 are actually > better. EasyProg from Olin is another example. Even though > there is only Olin's pc host software implementations now, > People have ported his EasyProg firmware to Wisp628 (EasyISP). > It is a pity that WinPIC800 is closed source but this is the > copyright owner's decision and they must have their reason > to do so. Now IIRC you've posted bootloaders (Chia I believe for the dsPIC). Now if you have a BCD and a bootloader, then exactly what does any of the three above bring to the table? I finally figured out that many believe that a programmer is an important tool in the development process because they use the programmer in the development process. Since I don't use the programmer in the development process, its functionality or reliability isn't of much relevance. When a programmer becomes a BCD, the overriding factors become cost and ease of construction. That's why the one chip, handful of resistors and transistor approach of the Trivial programmers are important. They become cheap and asy to build for BCD use. By the way I'll answer the bring to the table question above. The answer is that unfortunately Mchip has not seen fit to make their entire line self programmable. Personally I find it aggravating that parts like the 12F629 and 12F675 cannot be configured in bootloader configurations. That means for ICSP development that half your pins are being used as the interface for programming durin development. My interim solution to such problems is something along the lines of my PicDesigner project, which is a fully outfitted development box. Of course it has a bootloader interface so programming it is as simple as plugging it into a serial port. An obvious upgrade now would be to use a USB enabled 18F PIC and provide a USB interface. But in effect you use the big box as a bootloaded emulator, then when the project is finished use a BCD to dump the code into a smaller part for the final target. > Even those from Microchip are not necessary of high price. > And there are cheaper clones available. For a project or three, $30-$50 USD is expensive tooling for hobby use. As Wouter keeps pointing out, there are folks who are interested who have time >>> money. Students and tinkerers who can live with buying a couple of 555s and some discretes. It's one of the reasons that I'm sure that there is a bootloader market out there. When you get the PIC, you get the programmer too. It's just like getting a WISP628 without any of the associated board parts. Anyway I wrap up with my tagline: "PIC Programmers are overrated." BAJ -- http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist