login
Header Space

 
 

Programming Language Critiques

May 15, 2008 - 8:16am
Submitted by olecom on May 15, 2008 - 8:16am.

I wonder, what would be there:
TO DO: What's so special about string and text processing?
trace:
http://pdos.csail.mit.edu/~engler/cool-people.html
http://pdos.csail.mit.edu/~engler/
http://www.stanford.edu/~engler/
big Q: where is Open Source of all that cool stuff?

?Or is it again:

----
Indeed, one of my major complaints about the computer field is that
whereas Newton could say, "If I have seen a little farther than others,
it is because I have stood on the shoulders of giants," I am forced to
say, "Today we stand on each other's feet." Perhaps the central problem
we face in all of computer science is how we are to get to the situation
where we build on top of the work of others rather than redoing so much
of it in a trivially different way. Science is supposed to be
cumulative, not almost endless duplication of the same kind of things.

	R. W. Hamming
	One Man's View of Computer Science
	1968 Turing Award Lecture
----

I acually started form my very first topic here, which has some new appearance in lkml:

http://article.gmane.org/gmane.linux.kernel/680465

Andi says, patches will have feedback. I want: developing tools to have feedback too. Or it's just another handcrafting "know how" and "gimme the patch or stfu" of kernel developers. Oh, well...

I try to develop not just analysis, but text transforming. Document it, make it be an easy tool to access (e.g. `sh` && `sed` script). Allow it to be re-usable and re-editable for further similar tasks. Nobody understands, as it seems...
____

strange system...

May 15, 2008 - 12:03pm

So, after years of granted academia work, they went commercial, still having research going on.

Strange system...

All so-called 'papers' are available to download -- also strange for most for-paper-research workers...

I want every source code base to have tools (syntax highlightng, simple semantic annotations by designer's for easy static analysis, easy call-chain walkers, etc.) for editors, IDEs to prevent programmer from making bugs (obvious, or most known mis-uses of interfaces, or context-specific security holes known for decades, etc.) to do it as interactively as possible using syntax highlighting and non stupid user interface.

They have set up firm to provide payed bug-killling.

Where's the right and where is the wrong end of the telescope?

> to do it as interactively

May 15, 2008 - 1:55pm

> to do it as interactively as possible using syntax
> highlighting and non stupid user interface.

for example, one types spin_lock(&some_lock);, editor generates a checking template for all return to have spin_unlock(&same_lock).

Cases, when unlocking is done in another function, simple annotation can be used:

spin_lock(&some_lock); /* +ret */

/* this is bone-headed, inflexible crap: */
static void efi_call_phys_prelog(void) __acquires(efi_rt_lock)

What if many locks, functionality must be annotated? Gee...

As with sp@m, all can be done by `sed` easily. But yet another editor must be invented for such interactity.

Even more crazy. Use ordering of source as locking is done and preprocessor:

#       define    _some_lock_
        spin_lock(&some_lock);
..../* same body */
        spin_lock(&some_lock);
#       undef     _some_lock_
#####
#       define    _some_lock_ret
        spin_lock(&some_lock);

..../* unlock is in another body */
        spin_lock(&some_lock);
#       undef     _some_lock_ret

Path Projection for User-Centered Static Analysis Tools

May 15, 2008 - 11:59pm

http://www.cs.umd.edu/~jfoster/path-proj.pdf

To address this problem, we present a new user interface toolkit called Path Projection that helps users visualize, navigate, and understand program paths|call stacks, control flow paths, or data flow paths -- which are a common component of many static analysis tools' error reports.

Again, why not in editors? Why to do not allow errors in the first place, using various policy applications while editing source?
____

contrary, what D. Knuth said, Massalin dreamed

May 21, 2008 - 6:18am

After lwn's "Open Source Security Report".

Funny enough, that this "Coverity" thing and its autors are quite reverse, to what Knuth said:

authors these days often introduce arcane methods that outperform the simpler techniques only when the problem size exceeds the number of protons in the universe.

I.e. they have research on self-modifying code `C compiler, middle-ground (not a micro, not monolitic) way of building OS kernel -- quite interesting and IMHO useful. But they turned to the universe of bugs, most of which is invisible one (-- dark matter -- dream of modern corrupted physicist).

Year, bugs, bugs, bugs. Worth to note, that Massalins's hope for hardware support of self-modifying code in the future

It happens that the techniques are unusual, and make demands of the hardware that are not commonly made. But just as virtual memory proved to be a useful idea and all new processors now support memory management, one can expect that if Synthesis ideas prove to be useful, they too will be better supported.

was nicely "supported" by hardware bugs, which causes self-modifying code to fail.

Oh, well...

I wonder what all non-food-producing, non-oil-pumping, non-energy-generating people are doing, including never-ending kernel and compiler "development" with enterprize products for barin-washed-end-users.
____

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
speck-geostationary