| ??? 10/16/09 21:08 Read: times |
#169817 - But a large number of posters will guess Responding to: ???'s previous message |
Richard Erlacher said:
Then he should say so! A simple question such as, "What does <insert undecipherable text here> mean?" would do the job. Exactly! But is it a hw or a fw question? The answer will decide what hardware to use and/or what the source code lines will end up like. Richard said:
And you didn't really resolve the issue. Would a discussion about configurable features of a CAN controller be firmware or hardware? No single line of code is written yet, because the OP hasn't selected processor yet. But there is a potential candidate. And obviously no schematics either. So the poster can't fulfill your requirements for either a fw or hw thread. It will be fw that sets the values of the registers, but depending on the function of the hardware inside the processor, the amount of external hardware may differ a lot. If I could resolve it with a single post or two, it would have long been resolved. Further, it's not my place to do so. If he hasn't a question about a specific item in Hardware, Firmware, or Software, he shouldn't be asking US to hold his hand in making decisions about his task. Its up to him to decide on the task definitions, devise a solution, preliminary at first, and on that basis he can ask relevant and reasonably encapsulated questions. I see no purpose in helping someone solve a problem that starts with, "Why is there air?" I don't know who "he" is, since I am just trying to show the problem with putting a label on a subject that is more likely to have a random percentage in either side. Or are you of the view that no poster who don't already know the answer should post a question on this forum? Richard said:
Another example: If you see on the scope that a signal is sometimes 0V (or maybe 5V) and sometimes a fixed value in between - is it fw or hw? It very much sounds like two outputs colliding, but the reason may be that the user has a processor with push-pull support and have managed to configure a pin as an output when it should have been an input. It looks to me as though you've forgotten what an oscilloscope or an 805x port pin can do. Colliding outputs don't generally produce what will appear as a DC level on an oscilloscope. If he's seeing a DC level with a 'scope, he's already decided he has a hardware problem. It's either a misdesigned/miswired circuit, or a failure to understand how the hardware works. If firmware is an issue, that will surface later. Most of these issues are raised by people who are too lazy to get out their oscilloscope, or who haven't even got one on hand. Good answer. But an answer showing why it doesn't work well to separte fw or hw. If the user is good enough to solve the problem, there will be no need to post. And if the user do post, he/she will most probably not know if it is hw or fw. And the true classification will be deduced in the last chapter^H^H^H^H^H^H^Hpost - except of course that most posters don't return back with an explanation of what the error was. The thread just dies when the problem has been (or seems to have been) solved. Richard said:
"Does not require 805x" shouldn't be discussed here at all, except, perhaps, under special circumstances on the "CHAT" forum. If you want to discuss general computing issues, this is probably not the most fertile ground. Great progress. Your view is describing the current situation. "8052 Forum" for 805x-related threads and "Chat Board" for discussing Le Tour or similar. But suggesting changes requires suggesting changes or it would not be changes. That is why I am suggesting a "General Hardware" forum, for discussing how to charge NiMH, or peoples experience with RS-485 or if new designs should use 5V or 3.3V etc. So instead of telling me that "Does not require 805x" shouldn't be discussed here at all [...]" - how about telling what advantages/disadvantages you would see with a generic hw forum? Do you think it would change peoples view of 805x processors? Would it change the quality of posts in the "8052 Forum"? Would it waste bandwidth, making the server too slow? Remember that this is not a democracy, where we may use votes to introduce such a forum. It is still Craig Steiner who makes decisions, just as it is up to him to decide if we wants to read posts on the forum. All we can do, is to try to see if we can come up with suggestions that may possibly improve the site. Suggestions thaty may not have been relevant 10 years ago, but may possibly be relevant today. Richard said:
For a forum to be meaningful, it will need a critical mass. If too many people bleed off because they work with different processors, making "Chat Board" the only alternative, then it will not matter if people asks good or bad questions. That makes sense ... that's probably why there is a CHAT forum. However, THIS forum has had, and continues to have, that critical mass. Many new applications still use the 805x-core MCU's and many edu's throughout the world still teach courses with the 805x as its MCU-core. There are still MANY application boards sold with 805x MCU's as the processor. People who read and post here are interested in those, and not the 32-bitters. Yes, some may want to learn about the 32-bitters, but ... is 8052.COM where they should be looking? Don't focus on defending 805x chips or the current functionality of www.8052.com. There are far too many products with 805x chips for it to need anyone to defend it. Instead consider the number of posters. Exactly how close are you to the limit for a critical mass? What happens if 5 more regulars goes silent? Still critical mass? Richard said:
What needs to be avoided is the loss of focus brought about by introduction of topics relevant to large (32-bit) processor cores, in a thread presented to 8052.COM, since 8052 is an 8-bit core. So if the OP did post here about running PPP with a 8051 with very little RAM and connected to a GSM module, your view is that the OP should continue that task even if it kills them? When someone starts a thread about a problem that is not well suited for a 8051, should you not consider suggesting them to get another processor? And if someone visits a forum discussing large ARM11 chips and wants to implement a lamp timer - should not anyone on that forum tell them that an ARM11 with a full Linux is not well suited for the task. The problem isn't that people gets scared away from 8051 chips. The problem is that every forum regularly gets new visitors that shouldn't have been there in the first place. Exactly what is your preferred answer if someone here wants to know how to copy data between two high-speed (480Mbit) USB hosts? Would you really consider an 805x chip for the task? Richard said:
But let's focus a bit more on the actual topic of this thread. A professional will probably not start a thread here asking about a specific problem unless it might possibly be an errata or similar. So the people who do create threads here with a problem are exactly the people who do not know if a problem is in the hardware or the firmware. That will obviously result in a very large percentage of threads getting the wrong tag. What's important is where the O/P believes the problem lies. If he lacks an opinion, then he hasn't given it much thought, has he? Does it matter? Having an opinion isn't the same as being correct. So the problem is how big that percentage will be who manages to select the correct subforum. If you visit other forums, with a number of sub-forums, you would notice that the moderators have to spend a lot of time moving threads around. Or the forums doesn't have any active moderation and the threads ends up pretty much randomly. The Keil forum just have a selection for processor architecture. That still means that a couple of percent of posters manages to post ARM questions as C51 questions or mix up 8051 and 80251. If separating hw and fw, you may have to be prepared to get possibly more than 50% incorret tags just because the people who posts makes the selection based on assumptions and it isn't impossible that these assumptions buts huge amounts of blame on hw when the problem is with the fw. They don't have the tools (or knowledge to use them) to test their hardware but they "know" that they have read the datasheets and made sure that their code is impeccable. |
