Slashback: Rocketry, Pythonation, Scoffing

Slashback tonight brings a few followups to recent Slashdot postings on the fate of model rocketry in the new, hypercautious America; a few Python gatherings for those who prefer that language to Perl; and a response from Los Alamos to recent claims of lax security. Enjoy!

Besides which, it's the hidden cameras that matter. An anonymous reader adds this followup to the story posted last month about Wired reporter Noah Shachtman's account of sneaking into classified areas at Los Alamos national Laboratory.

"In an email message to all Los Alamos National Laboratory employees, Pete Nanos, the current Director of LANL, responded with information suggesting that the Wired reporter who thought he had broken in to a 'top secret area' had in fact just crossed a cattle fence:

'The Wired reporter clearly did not enter a Laboratory security area. The Laboratory encompasses more than 40 square miles. The security force protects important assets within those boundaries but cannot -- and does not -- protect every square foot of property. Based on the article, it appears the reporter crossed a barbed-wire cattle fence, not a fence that protects a Los Alamos security area.
There is a small security area with several buildings (roughly 400 feet by 400 feet) near the driveway entrance to TA-33. That area is surrounded by a seven-foot-high chain-link fence topped with three strands of barbed wire. A security guard is stationed inside that area seven days a week and 24 hours a day. Clearly, the reporter did not climb that fence.
There are several other buildings outside the security area that are locked for property protection interests. They have no security interests. There are several gates and fenced areas on the TA-33 site, which are there for safety access control, not security.
It's unlikely the reporter would be prosecuted for trespassing; the Laboratory does not have law enforcement authority to prosecute, and none of the proper authorities witnessed the trespass.'"

Perhaps we can have a celebrity deathmatch. hfastedge writes "Ok, now that 2 perl conferences have been mentioned, I've been brought over the edge. Python is a language that is just as old, and arguably better from: most importantly a uniform standard of readability (enforced by using whitespace to delimit blocks (instead of {}), by avoiding overuse of cryptic symbols, and by a culture that strives to keep innovations as "pythonic"), and a rich development community. Anyway, normally, there are Python events in Europe, and a trail at O'Reilly's OSCON. But now, there is a far cheaper event taking place on March 24-28 in Washington DC: http://python.org/pycon/.

Examples of Python in action: 0, 1, 2, 3, 4, 5, 6, 7"

Fly up go phhhhhwwwtttpffffff .... MyNameIsFred writes "Slashdot recently discussed whether anti-terrorism laws would destroy model rocketry. The government has ruled, and the message is clear, "When it comes to the hobby of model rocketry, size does matter. And in this case, the magic number is 62.5 grams. That's the largest amount of propellant a single model rocket engine can have in it and still be exempt from a new set of federal rules that will go into effect May 24." What does this mean for the the big guys in model rocketry, who use engines larger than this?"

Comments

Shoot to kill

[The Bungi]

Clearly, the reporter did not climb that fence. [...] It's unlikely the reporter would be prosecuted for trespassing; the Laboratory does not have law enforcement authority to prosecute, and none of the proper authorities witnessed the trespass

I was the impression (probably because of one of those feverish Discovery marathons I tend to engage in when I get tired of coding) that the nice folks who guard US installations that contain either nuclear weapons of nuclear materials are allowed under federal mandate to shoot to kill. In fact that's what the warning messages posted along the fences of those facilities read - "lethal force authorized" or some such.

If that's the case Mr. Wired there (let's uncover the government's stupidity, for liberals everywhere!) was lucky he just stepped on some cow dung, as opposed to getting a 5.56 round in the chest.

Python is not just an alternative to Perl.

[Frater 219]

I like Python. I also like Perl. I think it's rather silly that people read these two as mutually exclusive. Both languages have their strengths.

• Perl is fast. There's no two ways about it -- Perl's code base has pretty obviously been heavily optimized, and programs written in it tend to be amazingly fast, especially considering that they are being bytecode-compiled rather than compiled to native code.
• Python is more readable. Past a certain degree of complexity of operations, Perl code starts to look like noise; Python code doesn't. Perl mongers tout the expressiveness of their language, but Pythonistas note theirs' readability.
• Perl is amazing for quick 'n dirty text processing. I wouldn't think of using Python when what I want to do is translate among document formats. That is Perl's native land -- where it ably and easily supplanted earlier Unix tools such as sed and awk, and remains in the forefront.
• Python deals better with complicated data structures. Ever read the perllol manpage? When you start to deal with nested structures in Perl, you have to play silly buggers with references, and if you do the wrong thing, you get ARRAY(0x6590) instead of your data. Python copes sanely with complex expressions and never gives you nonsense like that.
• Both Python and Perl have lots of modules handy. CPAN is hugely impressive. So is the Python standard library.
• Both Python and Perl speak the Web's languages fluently. The fact that both are embeddable as Apache modules (mod_perl and mod_python) should say enough -- but both can also parse HTML, XML, or what-have-you.
• Perl is great on the command line. There are so many ideas in text transformation that can be expressed in a single line of Perl -- no need to comment it, it's just a one-liner! -- and that can be quickly put to use in ordinary systems administration via perl -e.
• Python is great on the interactive top-level. Need to debug a Python module? Start Python in interactive mode, import the module, and start introspecting.
• Perl has plenty of room for your own style. This is a language whose possible syntax is huger than anyone has bothered to describe. If you learn it intimately, you can say exactly what you want in a minimum of bytes. There's a reason Perl fans like one-liners: the language is perfect for them.
• Python makes modularity and object-orientation make sense. In Perl, OOP is kind of a bag on the side, an extra feature tossed in to make modules easier to use. In Java, OOP is a Soviet political officer constantly intruding in your work and making sure you comply with the Party's way of doing things. In Python, OOP is just the way things work: everything's an object and it just makes sense.

Me? I use bash for one-liners, Perl for ten-liners, and Python for thousand-liners.

Python's "vs" issue

[mattdm]

I find it very interesting that so many Python users are so bitter about Perl, and so antagonistic. In fact, it was one of the things that kept me from even bothering to look at Python for a long time -- all of the Python advocates come across like a bunch of jerks. When I did get around to looking, I found that it's a nice, interesting, and useful language -- that doesn't really compare to or compete with Perl in a meaningful way at all.

Two of Perl's main strengths are 1) CPAN and 2) regular expressions integrated naturally into the language. Python's libraries are pretty good, and there's a lot of good stuff out there, but with Perl, I can pretty much count on 99% of anything I want to do having been done already. And sure, Python can "do" regular expressions -- in approximately the same way that one can do them in C or Java, by making a series of function calls.

At least on the second of these points, Python isn't even in the same *business* as Perl. There's just flat out no meaningful comparision. Python has *a lot* of strengths, but they're totally different from Perl's. So why do Python advocates get so worked up about something their preferred language fundamentally isn't designed to do? Why don't they raise a big stink every time someone mentions Java? That seems like a more usefully-comparible application space. Or C++, for that matter.

Re:Python's "vs" issue

[smallpaul]

At least on the second of these points, Python isn't even in the same *business* as Perl. There's just flat out no meaningful comparision.

I find it kind of odd that you think that a language with regular expressions in the syntax and one without cannot in general be used to solve the same kinds of problems. Does typing the parens change the nature of the problem that much?

But more to the point, if Perl's raison d'etre is built-in regular expressions (as you claim) then why is it used for (e.g.) SlashCode. What about building a large weblog community is regular-expression-centric? The vast majority of sizable Perl programs have very little to do with regular expressions.

Python has *a lot* of strengths, but they're totally different from Perl's. So why do Python advocates get so worked up about something their preferred language fundamentally isn't designed to do?

There was a period in the mid-90s when it was very common to be forced to write Perl because there was just some Perl code around that had to be maintained. The liklihood of this happening to an individual is proportional to the popularity of Perl. If Perl ceases to be popular then there won't be that much Perl code laying around to be maintained. Just today I was scrounging around in some open source scruffy Perl code that could have been better written in Python. That's the argument from self-interest.

The argument from emotion is that the popularity of Perl rather offends people's sense of competition between technologies being on the basis of quality rather than luck or marketing...in somewhat of the same way that it is typical to be annoyed at the popularity of IIS on Windows 2000 Server. In a sense, Perl is symbolimatic for some people of the less admirable qualities of our industry. What if the "Perl Aesthetic" (which discounts the importance of cleanliness and orthogonality) was to become the norm in other areas of our business?

Why don't they raise a big stink every time someone mentions Java? That seems like a more usefully-comparible application space. Or C++, for that matter.

Well for one thing, Python interoperates really well with Java and C++. So every line of Java and C++ out there can be construed as bolstering the argument for adding Python to the mix a a glue language. Whereas it is very rare to take a large Perl system and say: "let's bolt on some Python to provide a high level interface to this thing." For another thing, the liklihood of Python replacing Java and C++ in the hearts and minds of corporate application developers is quite slim. (especially considering the performance issues) But Python really could replace Perl for everything but 1-liners.

But the most important reason Python users trash Perl is because the first question a newbie asks on hearing about Python, an object oriented scripting language is: "How is it different than Perl?" It is rare for people to ask how it compares to Java or C++ because typically they are expecting Python to complement or extend systems built in Java or C++. If there were a clear consensus out there about when to use Python and when to use Perl then there would be little anti-Perl sentiment in the Python community. But saying that "Perl has built-in regular expressions" is not the same as providing a guideline for when it is better to use Perl than Python or vice versa.

Re:Whitespace BAD, Mkay...

[Anonymous Coward]

i thought pretty much the same thing, until i actually tried writing some python. takes about 20 minutes of hacking out some code of your own, then you no longer notice the whole whitespace thing at all. give it a whirl - after all, you're already indenting your code properly, right? well, writing python is just like that, except you don't bother with the curly braces. that's the whole difference.

code-writing code is a special case not worth optimizing for. by far most of the code that actually matters (including, most especially, any and all code-writing code) is written by humans, so naturally languages should be optimized for readability to humans rather than ease of printing by machines. but that goes without saying; if it were ever otherwise, we'd all still be writing machine code directly.

and other than that one special case, your entire argument would seem to rest on... um, nothing whatsoever. except your own personal dislike, which you're welcome to and which is perfectly valid - but it doesn't prove anything to anybody other than yourself.

now, that's not to say this feature of python's is all good and wonderful. there's one big backdraw to it, which i can't make any excuses for - with no explicit block delimiters in the language, you can no longer use % to jump betwen block start/endpoints in vi. that, i do miss when coding python.

Re:pretty obvious, don't you think?

[Rocketboy]

The 62.5 gram restriction has been a fact of life for high-power rockets for years and we've learned to live with it under the former set of regulations. The current problem is that since BATF has classified the AP propellant as an explosive, anyone who touches it in the transport chain from manufacturer to end user has to have been trained, had a background check, etc. UPS, FedEx and the US Postal Service are *not* going to do that for tens of thousands of package handling and delivery people, any one of whom *might* touch a package of rocket motors at some time during their working lives. They just won't allow anyone to ship the stuff. We'll be able to legally make and use the stuff but we won't be able to ship it from the manufacturer to the end-user. This is not helpful. Worse, since AP won't explode and can't be made to explode no matter how hard you try (despite BATF's classification as an explosive,) we won't be any safer from 'terrorists' than we were before. Regulation for regulation's sake, not for any reasonable purpose.

I feel safer already. Incompetant nitwits.

Rocketboy

Actually....

[Fencepost]

Most Python users fall into one of three camps:

• Used Perl, tried Python, liked it better and switched
• Used Perl, tried Python, found it better for some things, use both
• Never used Perl, don't care about it

I left out "Used Python, tried Perl, switched" because Perl's better-known and there are fewer people who started with Python. Besides folks who did that arguably don't qualify as Python users anymore.

There's a small subset of folks (most of whom really need to get a life and a few who are probably doing it for entertainment reasons) who publicly get really worked up about the superiority of Python. Similarly, in the Linux community there's a similar group of folks who get really worked up about Ruby. Remarkably enough, in the Perl community there is (wait for it) a similar group of folks who get really worked up about Perl. Some other products/projects with their own little fanatical subgroups: vi, emacs, Macs, FreeBSD, OpenBSD (and don't confuse the two!) and probably skript-kiddie toolsets.

The common feature that most of these folks share is that as far as the rest of us are concerned they need to get a life. I have no doubt that there are other shared behaviors within these groups, but if I went into those it could seem that I was just being nasty.