Slashdot Mirror

2-3 years ago, I was facing this difficult choice for writing my thesis: LyX wasn't quite there yet, staroffice (still from stardiv at the time) looked good but wasn't quite there yet either, and everybody else in the lab (99.9% of the people) were complaining about all kinds of problems in office97, or from migration to one system to another (one guy managed to go nearly all the way through from 2 to 2000, but it took him a lot longer than anybody else to get his PhD ;)...

WYSIWIG is great for short documents... something you manage to write in a few minutes and can still handle the layout of.

Anything bigger than a few pages, a few dozens of cross-references to sections, equations, figures, citations and word pukes. It doesn't do it straight away, though... but slowly at first and giving-up more and more errors as the document grows.
Then when you want to print to another printer that isn't the one you wrote your document for, the layout and page breaks go all over the place. This Isn't Normal.

I remember having had this discussion on /. at the time and several people advised me to move to LaTeX.

Sure I was shit scared to do anything the size of a thesis in TeX... need to compiling documents before you can see them, limited xdvi viewer, no spell checker... all in all LaTeX isn't very appealing for the new user.

But think about it this way: A 200page document is quite a big project. If it were a big programming project, would you rather rely on a limited point and click tool somebody who doesn't understand shit about the stuff you're really doing, or would you rather do it yourself with a powerful language like c, c++... insert your favorite language here.

There you go! and you don't expect the learning curve to be easy either, do you?

So yes, it was quite a difficult move for me, but fortunately, there are good documents on the net... just grab a copy of epslatex.pdf from a CTAN mirror and The not so short introduction to LaTeX 2e.

The most amazing thing about LaTeX is very simple: It's Open. This means that any part of your document, you can generate yourself from your programs. Need to generate a table with figures? just do it.

The same thing goes to two other programs I extensively used: grace and xfig. Yes they have somewhat limited interfaces, but you can generate the data from your own programs, so who cares about the interface?! they have open and well documented formats, it's the only thing that should matter.

For spell checking, I used aspell, again, who cares about real-time error correcting when you can do it in one go near the end?

For the editor, I don't know what you usually use, I use vi (improved :) and it works great. Use whatever you want.

Okay, I probably should stop being a LaTeX zealot, just think about it. Okay, you wouldn't start writing c++ code to just rename a few files... that's why bash is here for. The same way, to quickly produce a dirty document, wysiwyg is handy... but anything bigger than a few lines of code and you'll start to feel limited if you stay in bash instead of going for c/c++... same with documents... And when the program you're using is Trully Open, then you don't depend on The Big Corps who don't give a shit about you, just your money...

---

gnuplot is insanely powerful and flexible, because it's scriptable and supports a ton of output formats. I don't know if it can do the shaded area between two curves bit, though. It also has some limited support for 3D plots, although if you're serious about 3D you should really look at IBM's OpenDX.

grace is also a good choice if you like GUI plotting tools, but I'm so used to gnuplot that grace seems awkward...

--Troy

• cooltext.com (no source and limited to the 'standard' text logos with gimp)
• A Tutorial for Perl Gimp Users, namely the section on the Perl Server.

• cooltext.com (no source and limited to the 'standard' text logos with gimp)
• A Tutorial for Perl Gimp Users, namely the section on the Perl Server.

According to an article here, teleportation over optical cable is not even close to possible.

I'm surprised that nobody has mentioned formal methods for specification, design and implementation. Back in the 60s & 70s people like Dijkstra declared that programming was a mathematical discipline. Check out Seven Myths of Formal Methods for a summary that dispels the standard myths. This is also really good. Formal methods won't solve all problems, but they're useful.

You know, nobody builds bridges without mathematics anymore.

What? Software is more complex than bridges? People have mentioned design. Design is about breaking down complexity. Some people are good at it, most are bad. A good understanding of mathematics helps.

Can anyone confirm this leak? The Times doesn't say who or where this leak came from, and the Weizmann Institute doesn't seem to have any information on it, although they do have alot on RSA and public key crypto in general. And I would think any breakthrough like this would be kept under tight wraps, Israel isn't exactly the most open nation when it comes to spy and security stuff like this. Could be a red herring thrown out to make their enemies feel a bit more insecure. I'll believe it when I see it. Maybe.
^. .^
( @ )
^. .^

The way quantum computers work, it would take the same amount of time to crack 512 bits as it would to crack 56 bits, or any other value.

See, quantum computers don't do things serially like standard computers. They perform their operations on the entire data set all at once. It doesn't matter if the data set has 1 item or 1 billion items, it takes the same amount of time.

This is known as superposition. I don't know a terrible lot about the theory, but you can find out more at The Center for Quantum Computing. This Quantum Computing Tutorial is difficult to understand if you haven't done at least a little comp sci, and the one at qubit.org is better for people who've never heard of quantum computing at all.
---

Here.
--

From what I read here, the big difference between the functionality of Dr. Shapiro's device and that of ribosomes (rRNA), mRNA and tRNA is that Dr. Shapiro's machine could be made to construct strands of nucleic acids (read: more mRNA) while the RNA's can only build protein chains. I think this is a pretty big difference, because the machine's output can also be its input (which you really really don't want in cells). The upshot is that it won't be healing you any time soon - it deals in nucleic acids, not proteins, and it's too small to have little robotic arms or something to sew things up. :)

I have to do some more reading of the paper yet but what will really be interesting is how he proposes to keep the thing operating reliably - even DNA transcription (an incredibly fast and reliable process, IMHO) is prone to the occasional mutation (read/write error). DNA deals with this using redundancy and graceful failure (eg: the last GU in the subject mRNA string doesn't do anything at all instead of coding a second head).

I'm not sure how useful this device will be in doing any real computing - it will take some more work to develop algorithms for what looks like a fundamentally wacky programming paradigm (Have you booted your Turing machine lately? Now, do it with chemicals!) and plus there is no mention of the speed of this device. Who knows, it could be prohibitively slow. What looks nifty is the implications in data storage; in the paper, Dr. Shapiro writes, "The trace polymer created during the computation represents past state changes and head movements, as well as the symbols that were "erased" from the tape during each transition, and as such has several important advantages. First, the trace polymer renders the computer reversible." He goes on to describe how this could also make data storage inherently more efficient. So, rather than looking at cures for cancer I think we should look at very small and very auditable data storages.

Don't take my word for it. Go /. the horse's mouth! Buried in Dr. Shapiro's web site at the Weizmann institute is some Supporting material for the lecture including the paper itself (RTF).

From what I read here, the big difference between the functionality of Dr. Shapiro's device and that of ribosomes (rRNA), mRNA and tRNA is that Dr. Shapiro's machine could be made to construct strands of nucleic acids (read: more mRNA) while the RNA's can only build protein chains. I think this is a pretty big difference, because the machine's output can also be its input (which you really really don't want in cells). The upshot is that it won't be healing you any time soon - it deals in nucleic acids, not proteins, and it's too small to have little robotic arms or something to sew things up. :)

I have to do some more reading of the paper yet but what will really be interesting is how he proposes to keep the thing operating reliably - even DNA transcription (an incredibly fast and reliable process, IMHO) is prone to the occasional mutation (read/write error). DNA deals with this using redundancy and graceful failure (eg: the last GU in the subject mRNA string doesn't do anything at all instead of coding a second head).

I'm not sure how useful this device will be in doing any real computing - it will take some more work to develop algorithms for what looks like a fundamentally wacky programming paradigm (Have you booted your Turing machine lately? Now, do it with chemicals!) and plus there is no mention of the speed of this device. Who knows, it could be prohibitively slow. What looks nifty is the implications in data storage; in the paper, Dr. Shapiro writes, "The trace polymer created during the computation represents past state changes and head movements, as well as the symbols that were "erased" from the tape during each transition, and as such has several important advantages. First, the trace polymer renders the computer reversible." He goes on to describe how this could also make data storage inherently more efficient. So, rather than looking at cures for cancer I think we should look at very small and very auditable data storages.

Don't take my word for it. Go /. the horse's mouth! Buried in Dr. Shapiro's web site at the Weizmann institute is some Supporting material for the lecture including the paper itself (RTF).