Is it the default now for individual entries on blogger.com blogs to be displayed as lightboxes, or is this just an unusual option this one user chose?
Just try reading it on an iPad. It's an absolutely miserable experience. It feels like they tried to reimplement scrolling, which, of course, they get totally wrong.
It baffles me that companies spend significant amounts of time and resources making a user experience that works great rather awful[1].
Oddly they are continuing to make it worse. For at least a year you have needed JavaScript enabled to view Blogger sites. It's a blog, why do you need JavaScript to display a loading screen! And they scroll horribly in mobile Safari too.
And now these light boxes. I don't get it.
Related to the BoldPort product though, I think electrical engineering world is world could use some better tools. Earlier in the week one of our electrical engineers spent a day helping to debug a new board. The problem ended up being a signal name was renamed in one place but not another. If the current EE design tools don't have refactoring support much less some sort of electrical "uninitialized variable" error, there must be room for improvement in these tools.
I can't respond to the article itself, since I can't read it, since I refuse to turn on JavaScript for random web sites. BTW HN itself works great without JavaScript.
To your point about "uninitialized variable" nets, if the net was truly a singleton it would show up as such in a netlist report. This capability existed in board design tools I used 30 years ago. What might have happened is a renamed net might have been split into two different nets. E.g. a net that connected 10 places became two nets, one to 8 places, another to 2 places. Not a singleton.
There are additional checks that netlist generation tools can do. E.g. does every net have a "driver" or "output" on it, or does it consist of all "inputs". Absence of driver can easily be flagged as an error. However, there is one wrinkle. Sometimes, nets are "busses" and are bidirectional. Each pin of the net can be either an input or an output at various times. That's harder to detect as an error.
There are many ways to avoid problems like this. Two common ones:
1) READ THE LOGS. I'd bet even money that your problem shows up in the logs as a singleton net or as a net without a driver.
2) SIMULATE. Board functional simulation has been possible for over 30 years; it's not always done. Did your engineer simulate his design? Why didn't his simulation catch the problem?
Sometimes, in board design (as in FPGA design) it's faster and certainly easier to just get 99% of the way there ahead of time and debug the final few errors on the board itself. That might be the case here, so if your company made this decision a priori, then you shouldn't be surprised if you need to spend some time debugging a few simple errors.
BTW that sort of sloppiness just wouldn't fly in chip design. Chips are usually far far far too complicated; extensive pre-tapeout simulation is de rigueur. And it takes weeks (if you have much money and a very good relationship with a fab) or months to get revised silicon. Much easier to do a few cuts and jumpers to a board than to do a silicon spin.
As to "refactoring support", most CAD systems I was familiar with didn't have an easy way to rename a net across multiple "pages" of a board design. Maybe the tools have gotten better. IMO that sort of refactoring (if done extensively) is a losing proposition anyway. You're much better off with careful attention to detail up front. You should be very careful renaming nets that cross schematic pages; you wouldn't have the problem if you carefully thought about the names before you created the schematic.
In summary, even the best tools won't help a sloppy designer.
> In summary, even the best tools won't help a sloppy designer.
Sloppy engineering is bad. But excusing shitty EDA tools by suggesting the engineer is sloppy is not much better. Most EDA tools are crap at DRC (static rules that only cover geometry and physical connections) and user experience. (Most engineers wouldn't have access to very expensive tools that might be better than some.) You talk about FPGAs (I'm an FPGA/Verilog expert); have you ever looked at the logs? It's getting progressively better, but wading through hundreds and thousands of warning and info messages is not exactly the most effective way to tell an engineer that something might be wrong. Sure, you can write scripts, but the tools are shit out-of-the-box in finding problems and communicating them to the engineer. This is one example, but they're equally bad at guiding the engineer to good design practices that reduces faults, as I think they should.
Circuit design tools are worse, and engineers must rely on their experience and keen eye for detail. Some engineers don't have one or both of those. Are they sloppy? Can't the tools be more intelligent to help?
I think that there is a lot to improve in this domain, and I'm trying to do that with PCBmodE. Blaming the engineer and falling into the "digital Stockholm syndrome" isn't the way forward here.
I've used EDA tools (both board design and chip design) for over 30 years, and in that time there's been one constant: what the designers want to accomplish is usually at (or beyond) the limit of what the tools and platforms are capable of. So yes the tools are "shitty" for what we want to accomplish tomorrow. They're superb compared to what we had yesteryear.
You might not like my "sloppy" choice of words. But IMO the difference between good engineers and mediocre ones is exactly what you called a "keen eye for detail".
I agree the tools could be, and should be, a lot better. It's a copout by the tool vendors when they produce "thousands of warning and info messages". And yet the good engineer must wade through those messages. That is part of the necessary "get on with it" attitude.
It displayed as the default "Microsoft Word" layout when I first viewed it (after enabling Javascript), but then, when I used the "back" keyboard shortcut to try and come back here, the page re-displayed in a lightbox. Fun.
