- Feature Articles
- CodeSOD
- Error'd
-
Forums
-
Other Articles
- Random Article
- Alex's Soapbox
- Announcements
- Best of…
- Best of Email
- Best of the Sidebar
- Bring Your Own Code
- Coded Smorgasbord
- Mandatory Fun Day
- Off Topic
- Representative Line
- News Roundup
- Editor's Soapbox
- Software on the Rocks
- Souvenir Potpourri
- Sponsor Post
- Tales from the Interview
- The Daily WTF: Live
- Virtudyne
Admin
I'd be the frist to name a register in one of our IP cores like that...
Admin
our CIO complained a few years back that everything we produced was rubbish...(that was not the most pleasant meeting, you can imagine)... the word he kept using was "brol", which means rubbish in our dialect...
So when a few months later they wanted to replace our old legacy system with something more shiny, they were looking for a name... as we're in the tourist business, i suggested "Booking and Reservation OnLine", which would make a great acronym.... but my entry was not withheld, without a reason given
Admin
Sure there is. Try putting +24V into an output of an 8255 parallel I/O chip. Not only does the chip die, but it makes an actual crack sound (louder than just a "pop", not loud enough to merit "bang") and blows a hole in the chip's package. And lets the magic smoke out, both literally (a little curl of smoke was seen) and metaphorically (the chip didn't work any more).
Admin
Be careful with that word "embedded", because it covers a multitude of sins, ranging from the sublime (e.g. TI Sensor Signal Processor - the model I worked on had 576 bits of addressable memory) to the ridiculous (e.g. a system with two 12-core Intel CPUs).
Admin
Are you by any chance from the Dutch speaking part of Belgium?
Admin
Somehow, at my current job, it's evolved to include strapping a massive PC with a 5080 GPU to a PLC. If things go well, I'll actually get back into real embedded and replace the PLC with a collection of smaller (and cheaper) MCUs.
Admin
Fortunately, it can also mean highest volume production, which is often "best documented".
Admin
Reminds me of the time I was working on a project with some IEEE 1394 cameras. Even though the connector is designed to enforce the correct orientation, this particular one was a little sloppy in that regard, and I somehow managed to plug it in backwards. That was an embarrassing day...
Admin
Bring an infrared camera - you're gonna need it!
Admin
Documentation is like sex. When it's good, it's fantastic. When it's bad it's still better than nothing.
If you're looking at datasheets for flagship SoCs, it's mostly non-existent - it's a document spread across hundreds of documents and application notes and various little notes that engineers write. It basically amounted to "here's the code to make it work, you figure it out".
Admin
Not necessarily. Bad documentation, like 12,000 pages of things like "0x4001 Flangwoodle Derezzer, bit 1 = Drobbleworp, bit 2 = Woosaki, bits 3-6 = reserved, bit 7 = never set this bit" (Broadcom) that's worse than nothing, because if you had nothing you'd know the only way to get anything done was to take the vendor's binary blob, but with useless docs it can take you months to arrive at the same conclusion.
Admin
does it show?
Admin
Incorrect double negative in a forum like this? !!Hmmmm...
Admin
Truly bad documentation is not an issue. Not even the slightest. Anyone with half a brain can quickly discern if documentation is truly crap and just throw it away.
No, the real problem is what I like to call Evil documentation. That's documentation that looks right, feels right and is right... but only mostly. Its always well written and concise, except that one time it isn't. Its contents are always correct... 9 out of 10 times. Its exhaustive and complete except those things you don't even know exist and thus can't tell are missing.
In short, it's good enough and professional looking enough that it lulls you into a sense of trust and complacency only to betray you at the worst possible moment.
Now that is some true evil.
Admin
I'll throw in documentation that tells you one thing and then, several dozen pages later, tells you that the previous information was actually wrong and that this is the correct information.
While I can appreciate that there's probably some document management standard being adhered to whereby new information goes at the end, but at least mark the incorrect information as incorrect and put where the correct information can be found.
No-one - unless they've been burned by this kind of nonsense in the past - reads the entire document from start to finish before beginning.
Admin
That's not great, yes. But at least it told you that the previous thing is wrong. So a search will pick that up. And reading start to finish will as well.
What's real nasty is when the docs just give you two or more conflicting pieces of information with no indication as to which one is the correct one. You just have to figure it out on your own.
Can you tell I've been reading bad docs for longer than most people here are alive? Because I've been reading bad docs for longer than most people here are alive. :)
Admin
"brol" is a dead giveaway :-)
Admin
And if you're lucky, you might get a Yocto layer for the BSP. And that's if you're lucky!
Admin
You've not been following the introduction of AI, have you?!
IMO, that's one of the main things we're all going to have to relearn: that merely LOOKING professional and well-written and plausible is no longer any indication of accuracy.
Admin
Only when all else fails, read the documentation...
Admin
Point taken.
Admin
I have worked on multiple processors that had errata guides that were a quarter the size of the "real" documentation, sometimes bigger. What's worse than the documentation not being correct? The documentation not being correct because the designers built it wrong.
Admin
Not only its documentation, also its support ticket system.
And that at about the same time when iPhones and data plans became a thing. iPhones that deliberately refused to support Adobe Flash.
So you would just have given your on-call staff a bit more flexibility by providing them with iPhones and data plans to allow for internet access; and Oracle did just completely destroy that plan and would tie those on-call people to their PCs again to be able to access Oracle support when needed.
I fortunately managed to leave Oracle (having come in via the Sun acquisition, where I had ended up after the MySQL acquisition) just in time before all that would have affected my role there ...