Linked by Hadrien Grasland on Fri 14th Jan 2011 14:58 UTC, submitted by Debjit
GNU, GPL, Open Source "Steve Chang, the Chairman of Trend Micro, has kicked up a controversy by claiming that open source software is inherently less secure. When talking about the security of smartphones, Chang claimed that the iPhone is more secure than Android because being an open-source platform, attackers know more about the underlying architecture."
Thread beginning with comment 458375
To view parent comment, click here.
To read all comments associated with this story, please click here.
RE[7]: So what code is secure?
by Neolander on Sun 16th Jan 2011 12:58 UTC in reply to "RE[6]: So what code is secure?"
Neolander
Member since:
2010-03-08

That was the main problem with the referred languages. For a language to be a successful systems programming language, it needs to be the official programming language for a successful operating system.

There's something which puzzles me in this conclusion. If I remember well, UNIX was not initially C-based, right ?

So why did Ritchie et al. decide to create C ? What was wrong with existing system programming languages on these days ? Why didn't they use the official programming language for a successful operating system instead of baking their own ?

Edited 2011-01-16 13:01 UTC

Reply Parent Score: 1

kerframil Member since:
2005-07-13

So why did Ritchie et al. decide to create C ? What was wrong with existing system programming languages on these days ? Why didn't they use the official programming language for a successful operating system instead of baking their own ?


Apparently, B - the successor to BCPL and predecessor of C - was an awkward fit for the PDP-11. Take a look at the section entitled "The problems of B" from Ritchie's historical treatise:

http://cm.bell-labs.com/cm/cs/who/dmr/chist.html

Reply Parent Score: 1

Neolander Member since:
2010-03-08

From this section, what I extract is...
-B handled strings like crap (indeed, for coding something like UNIX where just about everything is a string, this sounds like a big issue ^^)
-B didn't handle floating point data properly, due to its will to treat everything as a word
-B's pointer mechanism was slow
-B was much slower than Assembly as a whole

So it sounds like we have a mixture of language limitations (lack of control) and heavy performance issues, particularly in terms of pointer handling.

Wouldn't we have the same issues with "safe" languages like C# or Java nowadays ?

I mean, I have a hard time picturing myself a scheduler written in Java, as an example, with GC kicking in and freezing the computer in the middle of a task switch.

Edited 2011-01-16 20:32 UTC

Reply Parent Score: 1