Post by thread:[PIC]: PIC development STINKS!

I say that PIC needs to offer free C compilers for their product, like some of their competitors, if they want to be #1! It was my understanding that Microchip set out to differentiate itself from its competitors by offering free development environments for its MCU's (like MPLAB). But today's toys are bigger, smarter, faster, and often times the level of abstraction that is offered by C is the preferable manner of doing things. So why dosen't Microchip continue their initiative of offering real, usefull development tools (LIKE A C COMPILER) for free? Its competitors sure do! It is my bet that they get kickbacks from HIGH TECH not to do this! {Quote hidden}> On Wed, Jul 31, 2002 at 09:35:26AM +0200, Wouter van Ooijen wrote: > > > The 16F84 will be obsolete when I will no longer be able to > > > buy it off the > > > shelf in ones. Although I do not necessarily qualify as a > > > poor hobbyist. > > > > OK, so IMHO entry one in the PICmicro FAQ should read like this: > > > > NOTES FOR A BEGINNER: SELECTING YOUR FIRST PICMICRO > > > > 1. The 16c84, 16f84 and 16f84a are obsolete. Don't use them, unless you > > have a *very* good reason, which could for instance be: > > * you can not get hold of any other chip > > * you are building an exsiting design > > * your design needs to run at 5.5 .. 6.0 Volt (16x84 only) > > And we have a winner! Wouter knows the few reasons to pick a 16F84. > > > > > 2. The recommended my first PICmicro's are (copied from Byron's mail on > > the gnupic list): > > * The 16F628. The superior successor to the 16F84. Check out my page > > here: > > http://www.finitesite.com/d3jsys/16F628.html > > to see why it's the clear winner. > > > > * The 16F87X family. Whether small (16F872), medium (16F876), or large > > (16F877) packaging, these chips are packed with features. Since they > > are self programmable development becomes real easy. Finally they price > > out really well compared to the 16F84. > > > > * The 18FXXX family. The new kinds on the block. Faster, larger, and > > with a superior instruction set and architecture, this family will soon > > become the rulers of the PIC world. It's going to take some development > > though because programmers and compilers for the part are slowly coming on > > line. > > > > 3. Please note that PICmicro chips can be mail-ordered from various > > sources in small quantities. Just a few: > > * http://www.phanderson.com > > * http://www.crownhill.co.uk > > * http://www.dontronics.com > > * http://www.voti.nl/shop/order.html > > Thanks for adding some places. If one can live with a $25 US order or paying > shipping then I'd add http://www.digikey.com to the list too. > > It's an excellent idea Wouter. Glad you thought of it. > > BAJ > > -- > http://www.piclist.com hint: PICList Posts must start with ONE topic: > [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads > > -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

On Wed, 31 Jul 2002, Brandon Stewart wrote: > I say that PIC needs to offer free C compilers for their > product, like some of their competitors, if they want to > be #1! It was my understanding that Microchip set out > to differentiate itself from its competitors by offering > free development environments for its MCU's (like > MPLAB). But today's toys are bigger, smarter, faster, > and often times the level of abstraction that is offered > by C is the preferable manner of doing things. So why > dosen't Microchip continue their initiative of offering > real, usefull development tools (LIKE A C COMPILER) for > free? Its competitors sure do! It is my bet that they > get kickbacks from HIGH TECH not to do this! Who offers a free C compiler for their microcontrollers? If a C compiler is all you need then wait a while and SDCC will be ready for you. Meanwhile, you can use Jal, which I understand to be quite powerful. If you're doing professional development with PIC's then buying professional tools is such a small drop in the bucket that free tools aren't even worth considering. (This of course is changing as gpsim, gpasm and now, SDCC are serious alternatives. I'll even fo so far as to say that for assembly programming that gpsim/gpasm is superior in many aspects to MPLAB. I'm of course slightly biased...) Scott -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

>If you're doing professional development with PIC's then buying >professional tools is such a small drop in the bucket that free tools >aren't even worth considering. (This of course is changing as gpsim, I just want to add that if you're NOT doing professional work, then Microchip probably has little interest in your home projects and hobbies. For this reason, I consider their MPLAB (being a capable utility) EXTREMELY valuable and I appreciate Microchip providing it free of charge. Believe me, Microchip's not missing out on anything by not providing a free C compiler. As a hobbyist, you might WANT one, but that doesn't necessarily make it a good investment for Microchip to develop and hand out free of charge. --Andrew _________________________________________________________________ MSN Photos is the easiest way to share and print your photos: http://photos.msn.com/support/worldwide.aspx -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

> I say that PIC needs to offer free C compilers for their > product, like some of their competitors, if they want to > be #1! It was my understanding that Microchip set out > to differentiate itself from its competitors by offering > free development environments for its MCU's (like > MPLAB). But today's toys are bigger, smarter, faster, > and often times the level of abstraction that is offered > by C is the preferable manner of doing things. So why > dosen't Microchip continue their initiative of offering > real, usefull development tools (LIKE A C COMPILER) for > free? Its competitors sure do! It is my bet that they > get kickbacks from HIGH TECH not to do this! Whenever I heard someone telling (shouting?) that X should be provided (for free of course!) I get a bit uneasy feeling over me. If that someone realy feels that way, why not get a (free!) C compiler (free Eagle Light, free datasheet, free PC/Linux C compiler, freeDOS, free ...) and *write* that stuff you feel should be 'freely available'? BTW this remark applies for X = compiler, design, FAQ, (beginners) manual, etc. I'm sure James will gladly host whatever free stuff is produced! Wouter (probably also speaking for a few other free-stuff-providers) -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

On Wed, 31 Jul 2002 09:43:55 -0700, Scott Dattalo wrote: >On Wed, 31 Jul 2002, Brandon Stewart wrote: > >>I say that PIC needs to offer free C compilers for their product, >>like some of their competitors, if they want to be #1! It was my >>understanding that Microchip set out to differentiate itself from its >> competitors by offering free development environments for its MCU's >>(like MPLAB). But today's toys are bigger, smarter, faster, and >>often times the level of abstraction that is offered by C is the >>preferable manner of doing things. So why dosen't Microchip continue >> their initiative of offering real, usefull development tools (LIKE A >> C COMPILER) for free? Its competitors sure do! It is my bet that >>they get kickbacks from HIGH TECH not to do this! > >Who offers a free C compiler for their microcontrollers? TI does for the MSP430 flash family -- in fact a complete HARDWARE debugger, in-system programmer and excellent Windows IDE with C Compiler/Assembler (puts MPLAB to shame) can be had for $99. Neither Microchip's ICD or ICD2 comes close in functionality. There is also an Open Source (a GNU C port, I think ?) MSP430 compiler in development as well. Some really nice mid-priced professional 3rd party tools are emerging now also as well as low-cost tools that would suit a hobbyist. After completing a couple MSP430 based projects, I can tell you they are much much easier to work with than any of the PIC family parts (including the 18F parts). They are lower power, about equivalent speed, have a nice linear address space (ROM, RAM and I/O), a nice orthogonal *16 BIT* CPU core design and have a real stack. Oh yes, did I mention a contiguous block of RAM? These parts are really nice to program whether you are a "C guy" (like me) or an assembler whiz. They also are building a nice range of parts, from 2K-ish (around $1-$2 singles) to full featured parts with 60K Flash/2K RAM, multiple UARTS, PWM's, TIMER's, ADC's, LCD drivers, etc. On-board JTAG ports and hardware debug registers make hardware debugging an "always works" proposition, unlike the PIC's less than straighforward method. The MSP430's also have a built-in bootloader in ROM. Very nice. They are so dead simple to use, I'd actually recommend them to someone just starting with microcontrollers over the PIC. I really wish Microchip had scrapped the basic PIC architecture with the 18F series. They had a chance to dump the arcane PIC architecture and start from scratch ala Atmel or TI. For anything but very basic and simple tasks, I much prefer the Atmel or TI parts now and probably won't move many (if any) new designs to the 18F parts. I just can't see investing in all new tools for a part that's not much better than the more advanced 16F family. Anyway, I don't hate PIC's or anything and I doubt I'll ever quit designing with them. I just think it's a good idea to always evaluate alternatives and right now I see a few nice ones out there. Matt Pobursky Maximum Performance Systems -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

Andrew wrote: > Believe me, Microchip's not missing out on anything by not providing a free > C compiler. As a hobbyist, you might WANT one, but that doesn't necessarily > make it a good investment for Microchip to develop and hand out free of > charge. I strongly disagree on this point. I can name a number of very high volume products which have PICmicro MCUs built into them simply because an engineer wanted to try out some ideas before committing to a design. The MPLAB IDE being free of charge allows an engineer to develop an application and try it out in a circuit without having to go to his manager/VP/purchasing department to request the software. When it comes time to develop the product for real, why would he go with any other device? Even if the shop is committed to using other manufacturer's chips, I doubt they would change the design and rewrite the software. This scenario especially true in the current business environment. Nobody has cash for engineers to experiment or redesign products that already work. I would like to point out that HI Tech Software have made their "PICC Lite" compiler available free of charge (download it at http://www.htsoft.com). I've been playing around with it for a few months now - you can develop code for the PIC16F627 and PIC16F84 and use it within the MPLAB IDE. This is a really nice combination for developing PICmicro MCU applications and the one I'm recommending now to anybody that sends me an email asking what is the best way to get into the PICmicro MCU. myke -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

> -----Original Message----- > From: Brandon Stewart [SMTP:spam_OUTbrandonstewartTakeThisOuTATTBI.COM] > Sent: Wednesday, July 31, 2002 5:21 PM > To: .....PICLISTKILLspam@spam@MITVMA.MIT.EDU > Subject: [PIC]: PIC development STINKS! > > I say that PIC needs to offer free C compilers for their > product, like some of their competitors, if they want to > be #1! It was my understanding that Microchip set out > to differentiate itself from its competitors by offering > free development environments for its MCU's (like > MPLAB). But today's toys are bigger, smarter, faster, > and often times the level of abstraction that is offered > by C is the preferable manner of doing things. So why > dosen't Microchip continue their initiative of offering > real, usefull development tools (LIKE A C COMPILER) for > free? Its competitors sure do! It is my bet that they > get kickbacks from HIGH TECH not to do this! > Todays bigger, smarter and faster toys probably don't use PICs. I really can't see what you are shouting about, you are given a (mostly) functional development environment with a perfectly good assembler and simulator. If you are a professional who needs the advantages of a high level language, then you pay for the best you can buy and that cost will be trivial, even in the short term on a profitable product. If you are a hobbiest, then there are hobbiest C compilers available (CCS) and Scott's SDCC port is well on the way. However, to use any of these products to produce compact, fast and bug free code with the minium of debugging time I'd say that a working knowledge of assembly was almost mandatory. If you are really that desparate to program in C for free, perhaps you should look at some of the more mature devices such as the 8051 which already have open source compilers. Regards Mike -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

On Thu, 1 Aug 2002 01:28, you wrote: Hi all, <snipped some great stuff> > If you are really that desparate to program in C for free, perhaps you > should look at some of the more mature devices such as the 8051 which > already have open source compilers. http://www.gnupic.org/ Has lots of great resources for the cost concious ppl, yes it is mainly *nix based, but with programs like CGYwin (IIRC have not used winblows for *nix emulation since my servers moved from NT4 to FreeBSD (way back in '96) Atmel is slightly lacking on the *nix support, but is catching up :o) (I have just released version 0.10 of the AVR ASM highlighter for KATE in KDE 3.0.X, and this will be released with the next KDE release (KDE 3.1.0 soon) Alot of people are working in this area, but all of us (weather pic, AVR, 8051 ...) are banging our heads ATM as there is a massive user base, which is refusing to donate their spare time in the spirit of GNU / GPL etc and help out, helping is as simple as writing documentation, finding bugs maybe even a bit of development work if you are a C junkie ;o) we all want free, we all want good tools, some of us want to be free from M$ crap, but if ppl do not help, then nothing happens. How long will it be till people like Scott burn out, IT IS HAPPENING, look at the number of unfinished projects, then ask why, it is usually not that it was a bad idea or no one is interested, but more like no one helped, no one bothered to let the author know his / her work is apreciated ..... I could rant all day, but you get my drift, IF YOU WANT FREE TOOLS, SUPPORT THEM, ASK WHAT YOU CAN DO TO HELP !!! Regards, Kat. -- -------------------------------------------------------------------------------------------- K.A.Q. Electronics Software and Electronic Engineering Perth, Western Australia Phone +61 (0) 419 923 731 -------------------------------------------------------------------------------------------- -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

myke predko wrote: > I would like to point out that HI Tech Software have made their "PICC > Lite" compiler available free of charge (download it at > http://www.htsoft.com). I've been playing around with it for a few > months now - you can develop code for the PIC16F627 and PIC16F84 and > use it within the MPLAB IDE. This is a really nice combination for > developing PICmicro MCU applications and the one. > I'm recommending now to anybody that sends me an email asking what is > he best way to get into the PICmicro MCU. The very limited number of supported PIC types of this compiler was for me the reason NOT to choose for it. I needed a compiler with support for (at least) the 16F873. Therfore I selected the demo version of CC5X by Bengt Knudsen Data (see http://www.bknd.com/), which has similar features as far as I can see, but supports many low and medium range PICs. There is a separate compiler for the high range. Its limitations (max 1000 statements, reduced optimisation) may be a drawback for large programs, but for learning and experiments I think it's ideal. Even with 500 statements I could produce quite fancy programs. For larger projects you can choose: buy the full function compiler or use MPLINK to combine object files. So far I have only positive experience with support. Rob. -- Rob Hamerling, Vianen, NL phone +31-347-322822 homepage: http://www.robh.nl/ -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

On Thu, 1 Aug 2002, Katinka Mills wrote: > On Thu, 1 Aug 2002 01:28, you wrote: > > Hi all, > <snipped some great stuff> > > > If you are really that desparate to program in C for free, perhaps you > > should look at some of the more mature devices such as the 8051 which > > already have open source compilers. > > http://www.gnupic.org/ Which I hardly update - if anyone has recommendations for other references that should go here please let me know and I'll add them. {Quote hidden}> > Has lots of great resources for the cost concious ppl, yes it is mainly *nix > based, but with programs like CGYwin (IIRC have not used winblows for *nix > emulation since my servers moved from NT4 to FreeBSD (way back in '96) > > Atmel is slightly lacking on the *nix support, but is catching up :o) (I have > just released version 0.10 of the AVR ASM highlighter for KATE in KDE 3.0.X, > and this will be released with the next KDE release (KDE 3.1.0 soon) > > Alot of people are working in this area, but all of us (weather pic, AVR, 8051 > ...) are banging our heads ATM as there is a massive user base, which is > refusing to donate their spare time in the spirit of GNU / GPL etc and help > out, helping is as simple as writing documentation, finding bugs maybe even a > bit of development work if you are a C junkie ;o) we all want free, we all > want good tools, some of us want to be free from M$ crap, but if ppl do not > help, then nothing happens. How long will it be till people like Scott burn > out, IT IS HAPPENING, look at the number of unfinished projects, then ask > why, it is usually not that it was a bad idea or no one is interested, but > more like no one helped, no one bothered to let the author know his / her > work is apreciated ..... I could rant all day, but you get my drift, IF YOU > WANT FREE TOOLS, SUPPORT THEM, ASK WHAT YOU CAN DO TO HELP !!! While I don't think you implied this, I (Scott) don't currently have any unfinished projects, just several on-going ones with no clearly defined deliverables! (I only have three projects: gpsim, gpasm, and SDCC). Kat is mostly right. Like BAJ, I won't even address the religious aspects of Open Source tools. Doing so is just a waste of time. I can tell you though that my personal motivation for Open Source development resides in the challenge of solving a complex problem. I view Open Source development exactly the same way I view basic Research. There's information and there are ideas. Information is the collection of ideas over time. I use information "out there" to expand my ideas and in turn reciprocate by contributing my ideas back to the "information pool". Anyone can use this simple model to guage how they wish to contribute. So if I may, let me re-phase Kat's last sentence: If you want Free tools, look around until you find them. If you find them then use them. If you discover a way to make them better then either enhance them yourself or make suggestions to the authors. If you can't find the Free tool you want, then find something similar. Ask the authors if they'd be willing to morph the tool for your situation. If that doesn't work, then spend some money and buy a commercial tool if it's available. Unlike most Open Source advocates, I don't believe there's anything philosophically wrong in purchasing honest commercially developed code. Again, using the Research metaphor, one's ideas can be enhanced with closed source too. Scott -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

> Who offers a free C compiler for their microcontrollers? > > If a C compiler is all you need then wait a while and SDCC > will be ready for you. Meanwhile, you can use Jal, which I > understand to be quite powerful. If you're doing professional > development with PIC's then buying professional tools is such > a small drop in the bucket that free tools aren't even worth > considering. (This of course is changing as gpsim, gpasm and > now, SDCC are serious alternatives. I'll even fo so far as to > say that for assembly programming that gpsim/gpasm is > superior in many aspects to MPLAB. I'm of course slightly biased...) > > Scott Is SDCC even actively maintained? And do they even have decent support for PICs? There web page hasn't been updated since last year, and even then they only mention STARTING to add PIC support. Is there some other place you are looking for your information? -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

On Thu, 1 Aug 2002, Nick Veys wrote: > > Who offers a free C compiler for their microcontrollers? > > > > If a C compiler is all you need then wait a while and SDCC > > will be ready for you. Meanwhile, you can use Jal, which I > > understand to be quite powerful. If you're doing professional > > development with PIC's then buying professional tools is such > > a small drop in the bucket that free tools aren't even worth > > considering. (This of course is changing as gpsim, gpasm and > > now, SDCC are serious alternatives. I'll even fo so far as to > > say that for assembly programming that gpsim/gpasm is > > superior in many aspects to MPLAB. I'm of course slightly biased...) > > > > Scott > > Is SDCC even actively maintained? And do they even have decent support > for PICs? There web page hasn't been updated since last year, and even > then they only mention STARTING to add PIC support. Is there some other > place you are looking for your information? I can promise you that SDCC *IS* actively maintained. I'd suggest browsing the mailing list archives. This will give you the most recent information as to the status. Also, you may wish to browse the GNUPIC mailing list archives where there's also some discussion about SDCC. But in a nutshell: The SDCC PIC port supports the midrange PICs (e.g. 16F84, 16F628, 16F877, etc.). SDCC is an ANSI C compiler with extensions. (For example, you can declare bits or specify variables at specific addresses). The PIC port is nearly complete from a functional perspective. The only thing lacking is floating point support, division, and mod. From a system level, the only thing lacking is a linker. If you care to get an idea of what exactly is supported, you can examine the regression test files in CVS: http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/sdcc/sdcc/src/regression/ Scott -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

> Alot of people are working in this area, but all of us (weather pic, AVR, 8051 > ...) are banging our heads ATM as there is a massive user base, which is > refusing to donate their spare time in the spirit of GNU / GPL etc and help > out, helping is as simple as writing documentation, finding bugs maybe even a > bit of development work if you are a C junkie ;o) we all want free, we all > want good tools, some of us want to be free from M$ crap, but if ppl do not > help, then nothing happens. How long will it be till people like Scott burn > out, IT IS HAPPENING, look at the number of unfinished projects, then ask > why, it is usually not that it was a bad idea or no one is interested, but > more like no one helped, no one bothered to let the author know his / her > work is apreciated ..... I could rant all day, but you get my drift, IF YOU > WANT FREE TOOLS, SUPPORT THEM, ASK WHAT YOU CAN DO TO HELP !!! If you have to help, then they aren't free. You can just as well "help" to get a good tool developed by paying for a commercial product. In the end you get what you "pay" for either way. ***************************************************************** Embed Inc, embedded system specialists in Littleton Massachusetts (978) 742-9014, http://www.embedinc.com -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

On Thu, 1 Aug 2002 20:43, you wrote: > While I don't think you implied this, I (Scott) don't currently have any > unfinished projects, just several on-going ones with no clearly defined > deliverables! (I only have three projects: gpsim, gpasm, and SDCC). Hi all Sorry of you thought that wa aimed at you Scott, it most certainly was not, I am just getting frustrated by the piles of good work done that just ends (more often than not in a half finished program) becuase the author lost interest due to no (or little) feedback. Regards, Kat. -- -------------------------------------------------------------------------------------------- K.A.Q. Electronics Software and Electronic Engineering Perth, Western Australia Phone +61 (0) 419 923 731 -------------------------------------------------------------------------------------------- -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

On Thu, 1 Aug 2002 21:16, you wrote: > If you have to help, then they aren't free. You can just as well "help" to > get a good tool developed by paying for a commercial product. In the end > you get what you "pay" for either way. I agree Olin, but for a small business like mine, a little time spent writing a tool is not a lot of money, and I mean a little time, take the XML script I wrote for KATE to highlight the AVR, It took all of 2 hours from start to finish, and that was from someone who has never written a word of XML in her life :o). Sure some tasks like simulators etc are big and need 1000's of person hours, but if 1000 people each wrote a small module, it would take no more than 5 hours per person (assuming it would take one person 2000 hours). I pay for software that works and does what I need, but I also use free / reduced price software if it does the same job as $$$$$ software :o) Regards, Kat. -- -------------------------------------------------------------------------------------------- K.A.Q. Electronics Software and Electronic Engineering Perth, Western Australia Phone +61 (0) 419 923 731 -------------------------------------------------------------------------------------------- -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

On Thu, 1 Aug 2002, Wouter van Ooijen wrote: > > Sure some tasks like simulators etc are big and > > need 1000's of > > person hours, but if 1000 people each wrote a small module, > > it would take no > > more than 5 hours per person (assuming it would take one > > person 2000 hours). > > No way, unless it is trivial software. In my experience productivity is > more like the logarithmical than linear in the number of persons > involved. I agree and add that of those 1000 people, 10 of them are faster than the other 990 combined. Of those remaing 990, 900 aren't cabable of even starting a 2000 person hour project. Of the 90 remaining, 80 don't have the persistance to finish. If these statistics are valid (and I think they're conservative), then the failure rate for projects is on the order of 90%. I'm not trying to be critical here. Ask Andy Warren how efficient an individual programmer can be. -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

I'll back this up from a project manager's perspective... 10 people @ 1 hour each IS NOT EQUAL to 1 person @ 10 hours. It's somewhat the opposite of manufacturing-line specialization. Have any of you used XP (Xtreme Programming)? I've implemented it on a large project and got to try/evaluate the sub-concepts within the methodology. The biggest and probably only failure was pair programming -- with any combination of strong/strong, strong/weak, or weak/weak programmers, productivity was no where near that of individual programmers (though there were some other benefits such as the weak members learning techniques from the experts -- still some training would've done that as well). I'm sure that communication and the processes to support communication and documentation are a major factor for this, but I feel sure there are other reasons. Cheers, -Neil. Wouter van Ooijen scribbled: {Quote hidden}> > Sure some tasks like simulators etc are big and > > need 1000's of > > person hours, but if 1000 people each wrote a small module, > > it would take no > > more than 5 hours per person (assuming it would take one > > person 2000 hours). > > No way, unless it is trivial software. In my experience productivity is > more like the logarithmical than linear in the number of persons > involved. -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

----- Original Message ----- From: Wouter van Ooijen <wfKILLspamXS4ALL.NL> To: <.....PICLISTKILLspam.....MITVMA.MIT.EDU> Sent: Thursday, August 01, 2002 4:37 PM Subject: Re: [PIC]: PIC development STINKS! {Quote hidden}> > Sure some tasks like simulators etc are big and > > need 1000's of > > person hours, but if 1000 people each wrote a small module, > > it would take no > > more than 5 hours per person (assuming it would take one > > person 2000 hours). > > No way, unless it is trivial software. In my experience productivity is > more like the logarithmical than linear in the number of persons > involved. > > Wouter van Ooijen Actually in my experience the bigger the team the greater the chance that you end up with some members that generate negative productivity. The really good people end up chasing problems introduced by the really bad ones. On a few occasions I have personally discovered people actively sabotaging a project. Regards Sergio Masci -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 > Actually in my experience the bigger the team the greater the > chance that you end up with some members that generate negative > productivity. The really good people end up chasing problems > introduced by the really bad ones. On a few occasions I have > personally discovered people actively sabotaging a project. Yeah, apparently, there used to be some of that where I now work. The people causing it are all gone now, but the principle, as far as I've been told was that these people were directly sabotaging a project so that they could fix the fault they introduced, and thereby look really smart to the management. Yay them. Not the way I'd choose to look smart. "The intelligence of a crowd is the the intelligence of the dumbest person in the crowd divided by the number of people in the crowd." Lynch mobs are really stupid. - --Brendan -----BEGIN PGP SIGNATURE----- Version: PGPfreeware 6.5.8 for non-commercial use <http://www.pgp.com> iQA/AwUBPUnS7QVk8xtQuK+BEQKK+wCfb/EVTC4t9who1S8/RJvZbTQjZZgAoKD5 nnPdRnmVIE7PF9jlPuPZxUPq =fjJ+ -----END PGP SIGNATURE----- -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics