Computer Science

« first day (779 days earlier) ← previous day next day → last day (3662 days later) »

12:41 PM

@Raphael Were you using some sort of a small program to convert tikzpics to png? Or did the program do something else? (I just remember there was some program :p)

Raphael

@Juho github.com/akerbos/scripts/blob/master/tikz2png

Juho

yeah, looks familiar

I wonder if something in roughly the reversion direction would be nice

something like: have a GUI that allows you to draw a graph (click a spot, a vertex appears etc.), and then you can output that as a piece of tikz-code

I don't know how quick people are at producing graphs "manually", but that would speed it up for me

Raphael

There are some SVG->TikZ scripts, but I have no idea how well these work.

@Juho In my experience, not. When you have some experience with TikZ. :D

Q: How to export svg to tikz

I was wondering if anyone had managed exporting .svg files to TikZ code. I do not mean include an .svg file into a latex document, but rather to obtain the code of an image so you can then edit it with TikZ. I know that this has been done here: http://www.texample.net/tikz/examples/india-map/ u...

Juho

I pretty much always have to look check how to make even the simplest things like "how did I make this edge dashed" or "how did I set a vertex weight..."

thanks!

I'd assume converting really an arbitrary svg to tikz is difficult

something graph-specific would probably be easier to implement as well

Raphael

@Juho Yea, it becomes better with time. Some things you just memorize, but most importantly you learn to a) always have the TikZ doc open on a second screen and b) navigate efficiently in it. And, of course, c) ask on TeX - LaTeX before fiddling for hours.

@Juho If the Debian logo works, many things should.

There certainly are features of SVG that do not carry over, but simple things should work.

Juho

12:51 PM

Yeah, I tend to fall back on earlier pics and just modify them

Raphael

However, you should check out the PDF+LaTeX export of Inkscape.

Juho

I actually didn't even know Inkscape had such support...

Raphael

@Juho The things you learn with a looming paper deadline. :D

Mathematica plots --> SVG --> tinkering in Inkscape --> PDF+LaTeX --> fixing the LaTeX --> inclusion in the paper.

The process is ugly as hell, but workable.

Juho

hehe, I see

Raphael

It's nice in that you get your document font in the graphics and can change details in text (e.g. add proper math). The positioning is wicked, though, so you don't want to have to adjust the exported LaTeX by too much.

5 hours later…

Wandering Logic

5:27 PM

Meh: "This site is currently in read-only mode; we’ll return with full functionality soon."

I don't recall much down time for se.

I wonder how many "9"s they have?

They're back.

vzn

5:45 PM

generally very good uptime. one )( thing not to complain about.

ran into this one recently. weird. boole not behind boolean algebra?

Q: Finding a graph-theoretic representation of expressions in Boole's algebra

I just read "Boole's Algebra Isn't Boolean Algebra" by Theodore Halperin (behind a paywall here). I don't have a strong background in abstract algebra, so, frankly, the paper is a bit over my head but the gist of it is as follows: the algebra developed by Boole in the 19th century has some strang...

graph-theory logic boolean-algebra

Wandering Logic

Yeah. But neither my company or my university has access to that part of SpringerLink, so I can only read the first two pages.

The axiomitization used in intro digital design classes are often called "Huntington's axioms".

This isn't unique. For example: Maxwell's equations (in the vector form we use them today) are due to Heaviside.

vzn

interesting hadnt heard of huntington axioms.

Wandering Logic

Well "often" is an extreme exageration.

I taught digital design in the past.

vzn

ok. nice.

also the term "algebra" seems to have a lot of meanings and gone thru many incarnations.

Wandering Logic

They are called "Huntington's axioms" in one text book.

But not in most.

vzn

5:54 PM

anything new with you?

Wandering Logic

Not really. You?

vzn

still banging head on collatz, endless :(

wrote new blog or 2.

your profile says you were on team building 1 of 1st manycore processors. curious, what was your role?

Wandering Logic

That was the MIT Raw microprocessor, out of which came a half dozen dissertations. We all did a little of everything. I mostly worked on one of the compilers.

On the hardware side, I mainly influenced the network design.

Oh yeah, and the cache designs. (We tried to do some crazy hardware/software codesign stuff to keep the hardware cheaper.)

vzn

6:17 PM

was it a dual core?

multicore seems not to be taking off as much as earlier (now optimistic?) forecasts...

Wandering Logic

7:16 PM

@vzn: 16 core

taped out in 2002, powered on in early 2003.

but we didn't have cache coherence

The cores were not particularly interesting. 5-stage MIPS-like pipeline and ISA. Inorder, stall on miss

1 hour later…

vzn

8:25 PM

did each cpu have its own cache? are you saying each chips cache was not coherent with the others? how is that addressed? was that overcome in the compiler or what?

just to get a lot of cpus on single chip was a big breakthru back then so their individual sophistication was not as much an issue (and think thats the general idea, continuing)...

just found this in physics chat room

The Nature of Scientiﬁc Proof in the Age of Simulations Heng

1 hour later…

vzn

9:55 PM

....

lol found this on google+ programming group

codebabes

"philosophy"

« first day (779 days earlier) ← previous day next day → last day (3662 days later) »

all rooms