100 word minimum

Delta time

I was trying to write about the latest incremental and horrifyingly minor breakthrough we've had, but I keep running through signal assignments in my head in an effort to determine whether the design will work under certain circumstances. I think it will, but it all depends on delta time. And I should test it under those circumstances.

I should stop thinking about it.

Delta time is an interesting aspect of hardware definition language. When you change the value of a signal during a process, that value isn't updated until the end of the process. So I've got all these decision statements (if-then-else) and I'm trying to visualize what the values of the signals, changed in an early part of the process, are when the decision statements are being executed. The whole thing is supposed to detect when a signal changes faster than a certain rate, so with everything changing all the time and decisions being made based on how things change, it all gets quite confusing.

Anyhow, I'm confused.

I should stop thinking about it.

Right now.

(But I think I really got it working. This time.)

Color me surprised

Lost in a Good Book did not suck.

We (Oz and I both) were really disappointed with The Eyre Affair, which consisted of a slim plot liberally interspersed with lots of authorial skipping around and saying, "Look! I'm clever!" I pretty much thought that would be the only chance I'd ever give Jasper Fforde, but when I saw a remaindered copy of Lost in a Good Book for only US$4.98 (plus 10% off), we went ahead and got it because it was cheap and recyclable.

Skip ahead to this weekend. I'm working really hard on not doing homework and reading fluff is always my first choice for procrastination. This book certainly qualifies. I also found it to be not annoying and even to include cool bits. Does that sound like damning with faint praise? I don't mean for it to, because I can unequivocally say that reading this book is a great way to avoid doing homework.

Triple word score

North Twenty-sixth Street between Grace and Franklin

I can't imagine why anyone would be playing Scrabble in the middle of the road. This tile has been embedded between the pavers for months, years maybe.

This is the one block on Church Hill with the original spallstone paving. When the city comes and does utility work on the gas lines and water mains, they always have to put the stones back down or the community association goes on the warpath. Seriously, tomahawks and everything.

Actually, I'm lying. But the idea of a bunch of irate homeowners and preservationists descending on city hall with stone age-style weapons amuses me.

Meanwhile, north of Broad Street, someone wants to Stop Gentrification.

This block looks pretty good to me. The fewer crack houses, the better, I always say.

With apologies to A.C. Clarke

Any sufficiently advanced nanotechnology is indistinguishable from the blue stuff growing in my crisper.

(I slay me.)

The inductive method

Math is excellent for clearing the mind of whiny, first world angst. I've been working on my math homework for most of the day and I'm feeling much better, thank you. I have a tendency in my spots of downtime to wallow in whatever is bothering me (which is usually not that bad as bad things go).

Lately the whine du jour (or, rather, de la semaine) has been "nobody will hire me because my background is so odd and therefore my career will be in essentially the same place as it was seven years ago except that I have over US$25,000 in student loan debt and no translation clients anymore because all I do is homework."

This is actually quite scary and not entirely unwarranted, because my job hunt has thus far been characterized by a total lack of response to all my applications. But I am not thinking about that right now. I just spent hours on my discrete mathematics homework solving recurrence relations and doing proofs. Very soothing.

Another technique I find useful when I run out of homework and this particular whine gets deafening is to apply to yet another job opening. I did that today too.

Reverse polarity

We've done so many about faces that we're back where we started. Finally.

Our design doesn't fit on the development board we are using, even after we break it into two parts and try to put the parts on the two chips available on the board. Then we decide to switch to a different board, not that we had any other options, and made up the necessary file of pinouts for that board. (The synthesis tools use the pinout information to link the actual wires on the board with the signals that we describe in the hardware definition language.) Once that's done, we work on revising the logic to work with the quirks of the new board. Then when Dr. Smith is helping us with some "bad synchronous description" errors, he mentions that he's got a better board and why don't we use that instead?

His board really is better. So it's back to making another pinout file, because the pins are totally different. We find, however, that this board is so much better that its documentation is ten times as long as the documentation for the other boards we tried. Likewise the chip on the board has got documentation that's two hundred times longer than the documentation for the other chips. Because printing out three hundred pages of stuff will annihilate our print allowances for the semester, we have to root around in these massive pdfs to find the information we need, which is oddly sprinkled through these many, many pages instead of collected in a handy table, as it is for the inferior chips.

So call me Grouchy.

We then spend the rest of the day troubleshooting weird errors that turned out to result from how all the components that were active low (need a '0' to turn on) on the other boards are active high (need a '1') on the new board. Except for the ones that weren't. Of course, if we had reviewed the umpteen pages of documentation, we would have found this out beforehand.

Now that that's all resolved (for the moment), we can start trying to figure out why it doesn't work.

Bellwether

The first thing I hear on the radio this morning is the weather report. They're talking snow, like six inches. No way! I check the weather page and they are saying no such thing, just a little sleet maybe.

By the time I go in to class, I've forgotten about it, but my fellow students are buzzing. We sit in our seminar class (where showing up is 99% of everything) and consider the weather.

"It's not going to happen," I say.

"Oh yeah? But Pinocchio washed his car today," Ratso counters.

"You did?" I ask Pinocchio.

"Yeah. It looks really good now."

"How much did it cost?" Cali-boy asks a leading question. "Twenty dollars?"

"No! It was $16.99."

"You paid sixteen dollars for a car wash? You know, the Rainbow Wash is only like twenty cents," Ratso says.

"Yeah, but I didn't want to get out of my car."

"So that's the indicator of whether it's really going to snow? Pinocchio spends a lot of money on a car wash?" I ask.

"Yeah," Ratso says, as if I should have known.

Class begins and we subside. Since we sit in the far back corner of the auditorium, the others play with their phones and laptops during the lecture. Text messages fly. I notice that the young man in front of me has a girly picture on his phone. When I comment on this later, I am told "No, that's not porn, it's wallpaper. Completely different."

Ratso is surfing on his shoe phone (it's not actually a shoe phone, but it's almost as big as one; it's a sort of hand-sized mini-laptop). "Psst," he whispers to me and holds it up to display a satellite image of a massive storm front bearing down on Virginia. "See?"

Overheard

We are out at dinner.

At the next table, a little girl says with authority, "Six is the new five." And launches into a treatise on shoes.

No quantum leaping

I solved my time stamp problem. It was, as suspected, programmer error. The software that the Hamsters interact with has a structure that is both sequential and layered. I just needed to move the time stamping process to another layer.

Now, on to the next problem! We've been working on a "smart" synchronizer for the Hamsters and, now that we've got our logic designed to work with what the hardware is actually doing, as opposed to what we thought the hardware was doing, our logic will no longer fit on the CPLD we're using to implement it. Unfortunately (or perhaps not), streamlining the logic is probably not an option because our current design requires 142 macrocells (spaces on the chip where the logic is put) more than the CPLD has available and I'm not sure that we can design away 47% of the logic without designing away 100% of the functionality. Fortunately, the CPLD sits on a development board with a second and as yet untapped CPLD. Having the two chips work together will involve a host of juicy clocking problems.

Lucky us.

Time out of mind

Or maybe I'm just going nuts.

Dr. Science and his crew want to compare the Hamsters with the pre-Hamster system. In order to do that, they want to be able to take Hamster files and dump them into the same software they've been using to analyze data collected by the pre-Hamster system. All quite reasonable except that the Hamsters generate data in a different (better) format. Consequently we've been stuck with creating a program that will take Hamster data and convert it to the pre-Hamster format.

Which is easy, theoretically. Except for the slight matter of how no one actually knows what's in the files generated by the old system. Or at least, that's what you'd think given how difficult it's been to get that information out of them. In defense of Dr. Science et al., the old system was built twenty years ago and the people who developed the software are long gone.

Today I finally dug through all the files they've been sending and found some answers. Ratso's mostly got the conversion software working. Now it's a matter of having the Hamster system generate a little extra data for Ratso's software to add in to the file it generates. One of the things we need is the start time and end time of the data collection session. It only takes five minutes to add a feature to the Hamster system so it will write out a file with the requisite time stamps. We hook up the Hamsters, have them collect data for a few minutes, then check the time stamps.

Oddly enough, the end time is earlier than the start time.

I know that time travel is not a Hamster feature. (We won't mention that to Dr. Science, because he'll get nanotubes involved and we'll get the kind of spec creep that haunts my nightmares.)

Hunched behind a monitor, Ratso says, "I think we should just let them worry about it."

"Yeah." I leave the Hamsters to run overnight to see if this happens when the start and end times are hours apart instead of only minutes.

Reptilian glee

Entrance to the A.D. Williams Memorial Clinic and Laboratories
1201 East Marshall Street

We found various interesting architectural details around West Hospital, the "brick monstrosity" as Oz calls it. I guess the longer you look at it…

The A.D. Williams clinic is tucked around behind West Hospital and, although it appears to be a mere appendage thereto, was actually built a few years earlier. The cornerstone says 1936. According to the MCV timeline, the laboratory and outpatient clinic opened in 1938 with the construction having been funded by the Public Works Administration. The vertical pillar standing out in the sidewalk supports a skywalk that connects the clinic with a newer hospital building across the street. This building is also slated for destruction by the university.

I really like the snake. It looks very enthusiastic about that staff.

After I took these pictures, I started wondering why there is only one snake per staff rather than two, which is what I recall seeing on various other medical insignia. A few minutes of research resulted in a plethora of material on the subject. In a nutshell, the double serpent staff is associated with Hermes and has commercial and occult connotations (which might explain my health insurance policy), while the single serpent staff is the insignia of Asklepios, a Greek physician who ended up being deified (the whole son-of-Apollo deal) as the god of healing, with all kinds of spectacular cures to his credit.

One snake is better than two, then, but why does this one look so jolly?

Not your ordinary fence posts

Fence around the Egyptian Building
East Marshall and College Streets

This is some kind of record—two homework-free weekends in a row. With the extra time we decide to go take some pictures of neat architectural things downtown, particularly those facing imminent destruction. The Egyptian Building is on the National Register, so it is certainly not in danger, but I love photographing this building and its surroundings. I'm not the only one, either.

The Egyptian Building is an Egyptian Revival structure that was built in the mid-nineteenth century to house what is now the Medical College of Virginia. It is very much the Victorian idea of what ancient Egyptian stuff looked like and the fence is no exception. I wonder why they thought that mummy cases were appropriate dér for a place where one is learning the healing arts. Then again, this was the nineteenth century and they were still catching on to the idea of washing their hands before surgery, so maybe receptacles for dead people are in order.

I've walked by the mummy cases many times, but I think today is the first time that I notice the feet poking out the bottom.

Okay, so the Victorians put feet on all their furniture. But not human feet. Not human feet sticking out of coffins.

'1' for the road

Remember that '1' we were so frantic to get yesterday?

We can't use it.

The line drivers are not behaving as I had assumed they would. Regardless of whether the Hamsters are putting out a '1' or a '0', the line drivers which receive the signal from the Hamster and send it on to the next level are always at '1'. At first we thought the issue was the crossed wires we found on the PCB (courtesy of the last Hamster team), but upon further review of the line driver documentation (Whoo hoo! More documentation!), I noticed where it said "A and B will be high when the line is inactive." So when the Hamster sends out an unchanging '1', the next stage receives an unchanging '1'. However, when the Hamster sends out an unchanging '0', the next stage receives an unchanging '1'. Since the logic for the next level is looking for the change between '1' and '0' and back again, this is obviously not going to work.

The solution is simple. We have the Hamster send out pulses when it's ready to start collecting data and a '0' (in effect, a '1') when it's ready to stop. We write a pulse detecting process for the next level so it can know the difference.

The two-and-a-half hours we spent yesterday on getting that '1'? Well, the extra familiarity with the Hamster documentation certainly won't hurt…

Make me '1' with everything

Mountain Girl and I are working on the Hamsters. We've put together the hardware, modified the software, synthesized a control circuit onto a CPLD, and now we're trying to figure out why it doesn't work.

We knew it wasn't going to work. We were prepared for that.

The first order of the day is to trace the problems back to the One True Error, the source of the cascading series of problems that leaves our pile of wires and circuitry with the status of junk. We go wrong at first by assuming that when we program the Hamster to write a '1' out to the Port F Pin 5, which we must use because that's how it's wired on the printed circuit board (to do otherwise would involve mutilating the board and soldering wires onto scratched up traces), the Hamster does indeed write a '1' out to that pin. We used the WriteBit() function, after all. It should write out a bit, right?

Once we finally discard that assumption, we put a probe on the pin and find that we get no '1'. We use assembly language and write a '1' to the pin by force. No '1'. We double check our Port F initializations and initialize Every Single control register to make sure that the port is set up to do what we want.

Port F ignores us.

We make pointed comments while casting sidelong looks just over the microcontrollers.

"Gee. I wonder why there are no hammers in the lab."

"Bad Hamster, no kibble."

We read more documentation. At this point we have six browser windows open to the various documentation files for the Hamster microcontroller. Repeatedly we are informed that Port F can be configured for output and how to do it. Repeatedly none of this works. Based on some characteristics of the more special control registers, I become suspicious that this is a problem with the data not getting clocked out to the port ever. I think that the "pclk" that is the default is a peripheral clock that does not actually exist. (The more I think about it, the more certain I become. We should probably have examined the other timing options but we were getting tired. Okay, so I just rechecked the documentation because I live to read documentation and that is not the problem. I think that the reference to the conflict between Ports A and F is perhaps not as complete as it ought to be.)

At last we review what did work. Port F Pin 5 was used to send out a pulse-width modulation signal. We find that if we use the PWM() function, we can set the duty cycle to 100%, making the pulse high all the time and thereby giving us our '1'. When we want a '0' we set the duty cycle to 0%.

This works.

I don't like it.

The joys of oxidation

When you form a layer of oxide on a silicon wafer, the color of the wafer changes according to the thickness of the oxide. This property is useful for eyeballing the thickness of your oxide, and certainly for visually identifying which wafers have been oxidized. It's also been used to create chip art (There's more on—and more—chip art at the Chipworks Silicon Art Gallery and some downloadable chip art wallpaper at Florida State University's Molecular Expressions) and, of course, IC chips.

Today in microfabrication, the professor mentions the difficulty of getting the resist to adhere to the wafer and how sometimes they end up spending hours in the lab trying to get resist to stick on a wafer that's being patterned with an image and text to give to a university donor. Mostly because they use old wafers that have been sitting by the oven for years. "Maybe you've seen them lying around the lab? You'll all get to make one at the end of the class. You can put anything you want on it—"

"Hey!" The class perks up.

"—as long as it's not obscene."

"Aw, man!" The back row is disappointed. I wonder what they wanted to put on their wafers, but I think I'd rather not know.

So, what shall I put on my wafer? I'm thinking of something for Oz, along the lines of "My girlfriend went to engineering school for seven years and all I got was this lousy wafer" or maybe something in Japanese like a big 気 (ki). Oz says that since I'm so miserable in the clean room I ought to keep the wafer as a reward (to remember the experience by?), but I can't think of anything to put on it. Maybe a riff on Rene Magritte's La Trahison des Images: "Ceci n'est pas un wafer." Or maybe Kitty-chan, she's everywhere.

Encoding

On my daily rounds of favorite places, I visit Arts & Letters Daily, where today I found a link to this American Scientist article touching upon what it is to be human. The author's main topic is how having a sense of purpose is integral to humanity and one of the prime movers in the development of humans, which is not really a new idea, but is nicely stated here. The part that grabbed me was the statement of a problem of identity, that our constituent parts (atoms, cells, etc.) turn over periodically, but we still remain who we are. We are information encoded in a self-renewing biological storage medium, which is interesting to think about even if you're not into logic design.

The idea of how an entity's parts can all be changed, but the entity remain the same is paralleled in some of the practices of Shinto. I've been reading A Year in the Life of a Shinto Shrine (here's a review, the link to which I'm including because it was written by the father of a friend from when I lived in California and how's that for a small world?) and today I hit the bit where he mentions the tradition, not so widely practiced now, of periodically dismantling and reconstructing shrines. This "continues most noticeably at Ise Grand Shrine, where every twenty years the main shrine buildings are entirely rebuilt on an adjoining site. This tradition of tearing down and rebuilding has actually been instrumental in ensuring a continuity of form and ancient construction techniques that would otherwise have been lost."

Interesting how patterns repeat. And that I should find both these texts on the same day.

Barriers to entry

Origami Girl bounds into the lab where Mountain Girl and I are working on the Hamsters. Since she's EE and our programs have diverged, we don't have any classes together anymore, so it's Old Home Minute.

"Guys! Would you believe that I got into grad school without totally applying?" she says.

"How's that?" "Where?"

"UVA! I started filling out their online application, but I never finished and then I missed the deadline to apply for financial aid. So I figured that was that, but then I got an email telling me to finish my application. I worked on it a little more, but I was tired and I figured, hey, no money, so why? And then the other night at, like, 1:30 am I checked my email and it said 'Congratulations! You have been accepted…' and I thought, now which of my friends did this to me? But it was real!" Origami Girl demonstrates high frequency oscillation.

"Well, congratulations. Are you getting a fellowship or something?" I ask.

"I don't know. I'm going to the information session and we'll see. I'm looking for a job around here anyway." She shrugs.

Clarence Wyatt Transfer

East Broad Street and Oliver Hill Way
(just off I-95 in downtown Richmond)

As of today, this building is gone.

I took this photograph one year and five days ago. It's not a great photograph, I took it from the Exxon station on the opposite corner of the intersection while Oz was filling up his tank. I never did get around to taking a better one. I figured I had time. After all, the building's been there for over a hundred years.

Hah.

This is the northern end of Shockoe Bottom. This part of town is filled with nineteenth century warehouses and crisscrossed with train tracks and cobblestone streets. This building has (had) a long brick shed extending from its opposite side. You can see a tiny bit of it at the edge of the photograph, especially the part of the roof that is raised and glassed in on the sides to form a skylight that runs down the spine of the shed. The lamentably un-photographed (I thought I had time) shed has train tracks running along one side, and it is (was) pierced on both sides with large doors topped with basket-handle arches and built of darkened wood panels fit diagonally into the frames. I assume that back when that stretch of track was in use, the trains would pull up beside the shed and their goods would be unloaded into the shed to be placed in another train or loaded out the doors in the other side into wagons. The area around the buildings is paved with spallstones and is in use as a parking lot.

The building was still in use. Sometimes you'd see lights on in the shed or the office part.

Then the shed burned down one night last fall, back when I was still using a cane to get around, so no pictures of that either. A demolition crew has been knocking down the shed and wrapping up the bricks (bricks aren't manufactured in that size anymore, so these bricks can be sold to restorationists who are working on similarly aged buildings). Today they started on the office building. I'd been hoping that it might be saved since it was relatively less damaged in the fire.

This reminds me: I'd better photograph West Hospital, only three blocks from here, before the university knocks it down.

Homework-free!

I am ahead with my homework. I was working on it till late Friday night in order to get to that point, but as my reward I get a homework-free weekend. This may be the only one I get all semester, so I'm going to enjoy it by, ah, not doing homework. And then not doing homework some more. This wasn't something I could plan for, or else we might have been able to arrange to take a small road trip or think of something to do besides bake cookies, which we're doing tomorrow. But this is nice. I'm really looking forward to a homework-free existence.

No question unanswered

Unasked questions get answered too.

But if I am not present, then I don't have to hear the answer.

In theory.

Okay, not to get all philosophical, but we are taking a closer look at this commercial weather station that Ratso thinks we can back engineer. I thought that wiring up a couple sensors to a microcontroller would be sufficient, but Ratso works at a marine supply store and has spent too much time staring at the stock and getting ideas. Hence the fancy weather station and its soon to be voided warranty. Each of the sensors on this weather station outputs data in different forms. Each sensor has a wire. Getting the data out of the wires and into our microcontroller will not be as simple as hooking the wire up to the microcontroller.

"No, the Hamster wants digital data. It can only read in 1's and 0's. What's coming off the wire is a variable voltage. The Hamster can only interpret that as either a 1 or a 0," I explain to Ratso, who's just brought up the "hook the wire to the Hamster" idea.

"But it's a DC signal," he says.

"Yeah, but that's not digital. We need it in bits. We'll have to use an A to D converter. Once the Hamster reads in the bits, it can interpret them as the actual data. We'll need, like, maybe three bits for that data?" I wonder.

Ratso looks blank. "So we need a three-bit A to D converter? But it's a DC signal."

"Well, yeah, but… I don't know if there is such a thing. Maybe you should ask Dr. Smith?" I know what he's in for, but I say it anyway.

Ratso goes to Dr. Smith's office. Mountain Girl and I work on our part of the project, finishing up the logic for the FPGA part. Then we try to compile and start cleaning up compile time errors.

Time passes.

"I wonder what's taking Ratso so long," MG says.

"I guess he's getting lots of answers. Do you think I should go see? On the other hand, that'll just distract Dr. Smith and it'll take that much longer to get Ratso out of there." Besides, Dr. Smith might well free associate a bit more spec creep for us.

Time passes. We clean out synthesis errors, resolve all the warnings that pop up during the translation process (compiling to hardware has a lot of steps: compilation, translation, synthesis). Continents shift.

I say, "Maybe I should go in there and get Ratso. But—Dr. Smith will come up with more stuff for us to do. God, I'm turning into such a hider, but that's the only way we're going to keep this project under control!"

Time passes. MG and I discuss how the software for other parts of the system need to be modified to work with the hub we're devising. I check my watch. I have class in a while.

Ratso staggers up to the lab and leans on the doorjamb.

"You get an answer?" MG giggles.

"I got the life story. He says we need a twelve bit A to D converter so we get enough resolution. I'm supposed to go ahead and order samples." Ratso lurches over to his computer and goes to the website of a chip manufacturer. He pulls of a screen filled with a full listing of A to D products. "Where do I even begin? What are all these?"

"Why don't you go ask—Pinocchio? He's using a twelve bit one in his project," I say.

Ratso mutters, "We've got sixteen bit ones upstairs. Why don't we just use those? But—oh, no—that'd be overkill, he says."

Know the difference

Oz says, "Yeah, I tell them at work how you rip on the IT people."

Indignant, I say, "I do not. I rip on the computer science people. There are two different programs. The IT people learn how to do things and the computer science people are too slack to study either IT or engineering."

"Well, somebody has to know how things work," he says and starts talking about read heads and software design.

Okay, so computer science people learn some of the stuff that computer engineers do. "But computer engineers are cuter. And they all have better skin."

"And they stand up straighter?"

I nod. "They do. And, based on an admittedly very small sample, I have to say that computer science students read Robert Jordan, while computer engineering students read Umberto Eco." Of course, in all fairness, I should say that if you remove me from that sample, you'd have to say that computer engineering students don't read at all.

Suits me

Or not.

Today is the career fair at my school. I've been getting prepared for ages, it seems. I get the suit, the shoes, decide on accessories, acquire appropriate foundation garments, have a tailor alter the suit slightly so it fits better, and on and on and on. Girl clothes have way too many rules. This morning I get up bright and early and get around to printing up some resumes, distribution of said resumes being the whole point of the suit and shoes and on and on and on. Adjust a comma here, update a little there, and ages later I have a folder of resumes. Then I dress and I'm ready to go on time, except I can't find a lipstick. I didn't bother to buy a new lipstick because I was pretty sure I had some old, petrified ones in the linen cabinet where I keep my medicine cabinet overflow. (Yeah, nasty old lipstick, but if you put chapstick on first, the lipstick goes on okay.) But I couldn't find one. "Maybe it's in one of those purses?" I run around the house looking into the little old purses I only use on the one or two times a year that I have to wear high heels and no lipstick there either. Back to the linen cabinet where I finally find a lipstick jammed beneath a q-tip box and unidentified toiletries of archeological significance.

I get to school without further mishap and run into Mountain Girl, wearing her black interview suit, outside the engineering building.

"Hey! We match," she calls.

More or less. Her shirt is cream colored and mine is electric, LED green (I'm trying to send subliminal messages to the recruiters), but otherwise, yes, we match. We walk over to the Commons building where the career fair is being held and up to the check-in desk, bypassing the employer check-in area, much to the surprise of the people working that desk.

"No, we are students," we explain. We only look like we have jobs.

"Way to go, ladies!" the employer check-in people say.

After we check in, we run into Cruise Ship Guy (who is wearing a name tag, so now I finally know his name!). He is all suited up and when MG compliments him even though he doesn't match us, he says how he got the suit at Goodwill for US$20 and tailored it himself. (I don't know why, he's got suits already, but anyway…)

"You know how to do so many things," MG admires.

"Well, I am from Jamaica. In my country we have to be able to do everything," he says.

Once they let us into the ballroom, we schmooze around with the three or four employers who are looking for engineers. Actually, they are looking for interns more than they are for full time employees, which is kind of scary for MG and me, but less so for CSG who still needs to get an internship in order to graduate. Also unnerving is that we have to explain what a computer engineering major is to recruiters who were ostensibly looking for computer engineers. Still, we get through it and even pick up a few leads.

Can employment be far behind?

Buffy Studies

I've never claimed to have my finger on the pulse of pop culture. I tend to avoid touching it because you never know what you might catch. But after enduring all the hype that surrounded The End of Buffy, whenever that was, and seeing TV shows come out on DVD, it seemed, if not exactly inevitable, but, well, if it sucks we just send it back. This DVD-thingy is a good way to watch TV shows, although we tend to yell "Commercial!" at every fade to black.

Oh, and it's addictive, like brain candy.

Oz points out, "The episodes are only 45 minutes long."

"Yeah, but you can't watch just one."

Thanks to lo the many hours of indoctrination in my youth, I really like the Young Adult (YA) story structure as it works in novels. (I was the girl whose parents limited the number of books she could check out of the library at once by refusing to help carry them. I went through fifteen or sixteen books a week.) They utilized the YA genre fiction structure perfectly for this show. If you don't know what I mean, read something by Diana Wynne Jones and you will. Anyway, all through engineering school I've been resting my brain with YA novels and now it seems that I can partake in video form.

I'll still get all my homework done. I can stop watching any time.

Learning the bugs

We have this homework assignment in Digital Design, another multi-hour engineering problem: design an 8-bit ripple-carry adder. The problem with the problem, as it were, is not with the design of the 8-bit adder. We've done little adders in vhdl before and I just used one of those, modified slightly to meet the specification we were given, because why write code when you can copy/paste? I got that done on Sunday. The problem is that we're using two different sets of tools (software packages) to translate the adder code into actual hardware, analyzing the gate delays, and then abusing our little adder with operations that will strain the timing constraints of the hardware. In English: We make it do hard math that takes a long time to see how long it really takes. Anyway, that isn't even the problem. The problem is that the tools are buggy! Yes, buggy! Buggy in the really charming "that only works if you compile at the command line because if you compile in the GUI you get an error" kind of way. Buggy in the "the signal names get changed in the timing constraints file so you have to go in and edit them by hand to match the post-synthesis structure file" way. So, between the bugs and my lack of familiarity with the software, I spent the day (yes, the whole day) synthesizing and simulating the adder over and over and asking the professor for help interpreting the mysterious error messages. I've barely gotten to the adder-abuse part.

The scary thing is that I'm the only one in the class who's started. It's due Thursday.

Opened eyes

Saturday night, really Sunday morning, and hours after I finish my discrete mathematics assignment, I lie in bed and prepare to drift off.

I'm having a little trouble drifting. Suddenly, homework—

"Oh. Is that what I did wrong? Did I misread the directions? Noooo." For one problem I got a different answer than what was in the back of the book and I didn't see how they got what they did. Now I do. Maybe.

"How tricky! Assuming that I'm remembering it right. Am I?" I scroll through the scrap of code in my mind. We were supposed to determine the Order of a looped structure. I did that problem several times before I got it "right" and then I was still wrong.

"No, maybe it was a different one. But, hmm. That's how they got an Order of 1. You enter the loop once and then leave without looping. Sheesh. And I spent so much time on it. Unless—"

Oz asks me if I need to get up and make notes.

"No, I'll remember this."

My subconscious never takes a holiday.

Oz makes a hat

I have homework. I can't procrastinate any more, so there's nothing for it but to sit indoors on this springlike (Setsubun strikes!) sunny day and work on discrete mathematics. Big O notation—so much fun and I'm sure it will be really useful. Now that I've committed those words to electrons, I'm sure they'll be haunting me shortly.

Oz amuses himself. Of late he has been disappointed with the fleece hats on offer at Target or wherever. He decided to venture into millinery and thus acquired navy blue fleece, needle, thread. Today is hat day since he ends up having a few hours (my hoped for two minute problems are liberally interspersed with the 45 minute kind). He cuts a strip of fleece, sews the ends together, sews closed one end of the resultant tube, and sews together the two corners of the sewn together end. He turns it inside out and, voila, hat.

"How do you like my hat?" The hat looks okay. Dark blue, hat-like, trapezoidal at the top.

"It's nice. You look, uh, Canadian. All you need is a beer, a string going through a hole in the ice, and an opinion on hockey. Eh?"

Setsubun

The first day of spring? By the Japanese calendar. Today and yesterday are Setsubun.

It's not very springlike on 3 or 4 February and, really, the deal with the groundhog makes a little more sense at this time of year. But I like the idea. Throw soybeans at the demons to scare them away. Shout at them and ask good fortune in. We could maybe give it a shot, except that all I've got in the house are canned beans. Oz would have to dress as the demon and it would be quite the mess. On the other hand, the cats would totally freak so it might be worth the cleanup.

I'm tempted to throw beans around the lab and call down good luck on all the senior projects.

Finite states

Procrastination is not too productive, but eventually the guilt kicks in and I do something. Not what I'm supposed to do (homework), but something I should do (submit more resumes to some really awesome jobs that I'm not going to get because of my minimal qualifications). I dither around and find a manga publisher requesting resumes from freelance translators. And consider the cover letter I'd have to write in order to convince them that a patent translator/computer engineer would be the ideal person to hire to translate—what genre? Shojo? Actually, I'd probably do really well with SF, as I can flim-flam improbable science with the best of 'em.

I also watch a video. Oz got Netflix and gave me his account information so I can pick videos to my heart's content. Our queue probably confuses the recommendation engine. So anyway, I watch All About Eve and, let me just say, Bette Davis is fucking awesome. Yes, my pop culture commentary has, like, a half century lag to it. On a more current note, this film does make about the best case against botox I can imagine. I mean, aside from the fact that the whole idea of injecting poison into your facial muscles to paralyze them is just nasty in so many ways.

DIY Engineering

Is there any other kind?

Mountain Girl and I are lurching on ahead with our part of the group project. We've figured out how to wire everything up and (supposedly) Dr. Science is acquiring the line drivers we need. If we don't hear anything by the end of the week, we're having Dr. Flight order them for us. Line drivers are the chips we use to convert signals between the RJ45 signals (Ethernet) and the low voltage pulses that can be handled by the CPLD, which we're using as a controller. Even without the line drivers, we can get started with the controller logic.

Our next step is to create the file that informs the hardware definition language of the CPLD pinouts (the pins of the CPLD chip are wired to various I/O devices like LEDs and switches and to the open pins through which the CPLD will send signals to the Hamsters). In our previous classes, this file was always provided to us but now we're sort of on our own. Because the CPLD is different and therefore has different pinouts from the devices we used before, we can't simply copy over the file we used before. Not without changing all the pin numbers anyway.

First we try to scavenge a UCF file off the Internet. We aren't the first people to do this and perhaps our predecessors have posted a file online. No joy. Oh well. Doing things yourself is good for you.

We take the schematics for the CPLD prototyping board and the schematics for the I/O board. These two boards are connected with two 40 pin sockets. Matching up the pin for LED 1 with the corresponding pin on the CPLD would be a simple matter if Pin 1 on the header of the CPLD board matched Pin 1 on the socket of the I/O board. Naturally, the pins don't match. Pin 1 matches Pin 39, Pin 2 matches Pin 40, Pin 3 matches Pin 37, and so on for forty pins times two. So say you're trying to hook the CPLD to LED 1. On the schematic of the I/O board, you find the I/O board pin that's connected to LED 1 and match it with the corresponding pin on the CPLD board, then you look at the pinout chart for the CPLD (which has 144 pins so big chart), match the connector pin to the CPLD pin, and write down the CPLD pin number for LED 1. Phew. We have plenty of devices to hook up: eight LEDs, eight switches, five buttons, and a seven segment display which is run with twelve wires. With two pairs of eyes and two heads, this is not too bad. I'm glad I didn't have to do this on my own, because I'd have gotten everything upside down and backwards before I got it right.

After we do all that, we use one of our old UCF files and just change the pin numbers. Then we surf around and look at pictures of groundhogs. "Ooh! This one's cute!" "That one looks like a prairie dog." "Eew. That one has yucky yellow teeth!" "Do you think they should do dental hygiene before they photo-op him?"

I notice that the grad student in the corner keeps shooting us the "My God! Will you people shut up? I can't concentrate!" look. He's going to have a really long semester.

Clean room

Lab in the clean room is not only dangerous, but also uncomfortable and dull. We have a Class 1000 clean room (no more than 1000 particles per cubic foot of air) and our protocol isn't as strict as at a commercial plant, where they would have a Class 10 or Class 1 clean room. I'm amazed that people at the big plants can stand it for any length of time. Then again, if they want out, all they have to do is break protocol and they're fired on the spot.

Protocol is the set of procedures one must follow in order to keep particles out of the clean room: bunny suits, hair nets, hoods, face shields, air showers, sticky mats at the doors to pull particles off the soles of your shoes. You can't take a pencil into the clean room, because a pencil works by rubbing particles onto paper. You have to use special paper. You have to use approved pens. If there is too much sodium in the ink, the particles released when you pull the cap off the pen will put sodium into the wafers, which will then have their electrical properties compromised by having too many ions.

Our protocol is sticky mats, booties, hair nets, safety glasses, latex gloves, and jump suits. Despite reality, it is assumed that all persons using the clean room are giants, so I get stuck with a huge jump suit, booties that flap around the soles of my shoes, safety glasses that rest painfully halfway down my nose (I have to use the big safety glasses to fit on over regular glasses since we aren't allowed to wear contact lenses), and big wrinkly gloves with squeaky air bubbles. I wonder if the petite researchers have their own personal gear that fits.

Many of the guys in my lab section (I'm the only one without the Y chromosome) are pretty big, however, and so I find myself spending most of the lab period looking at their backs as we gather around the rinse bench to wash our wafers. The washing process consists of dipping a basket of wafers in various toxic baths and standing around for several minutes while they soak. There's really only room for two people to work at the bench and I don't mind letting the foolhardy youths pour and mix the toxic chemicals because I've had more than enough physical trauma in the past six months, thankyouverymuch. The two boys who mess with the HF, hydrogen peroxide, and so on have to wear, over their protocol gear, face shields, rubber aprons, and big pink rubber gloves.

The rinse bench is back in the section of the clean room with yellow lights. This is where the resist is put on the wafers, and the wafers are exposed and developed (sort of like a photographic process, but not). The resist is light sensitive. You put resist on the wafer, expose the wafer to a certain pattern, the resist is cured in that pattern, you rinse off the uncured resist, do things to the wafer, then take off the rest of the resist so that the pattern of the things you did remains on the wafer. Lather, rinse, repeat. Eventually you get microchips.

Another particle attenuation method is air circulation. The air turns over in the clean room six times per minute and is filtered. Vertical laminar flow, in which air comes straight down through the ceiling and is vented out around the base of the walls, pulls particles down and out of the room. The ceiling is translucent and the fluorescent light is diffused so you don't get much in the way of shadows. Except in the yellow light area, the light is really bright and white. The fans make a lot of noise too. Between the weird light, the white noise, and the protocol gear, it's kind of a sensory deprivation experience.

Eventually our wafers are clean and dry. They are transferred back to their cassette for safekeeping till next week.