The Problem with Open Source is it’s Open Source
reynoldsftw.com
reynoldsftw.com
Some 'closed source' software is total crap too, but there we say 'product x sucks'. Here the fact that something is open source is taken to imply that all open source is dangerous simply because it is open.
It's a veiled way to promote security by obscurity. If you can't see the leaks we'll all pretend they don't exist.
Anyway, you're the last person on the planet I should argue that with :)
* Sendmail
* BIND
* CMU SNMP
* WordPress
All of these are open source.
Also, while the list would clearly be headed by Microsoft Windows, commercial software can do something open source can't. Pre XP SP2, Windows was a debacle. Then Microsoft spent 8-9 figures on security, and within 4 years, MSFT led the industry in code security. Sendmail is still a bad bet.
Again, don't get me wrong. If you have a choice between a random open source project and a random commercial project, and each is over some reasonable threshold of popularity, you're better off with the open source project almost every time. If you're a closed-source vendor and you aren't Microsoft, chances are your code is quite poor. It is at least harder to hide sheer awfulness in open source code.
But in any case, the reason I commented was, we can't cop out on the WordPress security debate by saying "it's open source, so the standards are different". WordPress is the Sendmail of our times. Lots of pro-bono effort went into it, but that can't obscure the fact of its very poor security design.
In the same posting mentioning BIND and Sendmail that had me chuckling here ;)
I hesitate to go so far as to call this post a troll, but the author is making very broad statements based on a very small sampling.
See recent twitter XSS attack, SQL sanitization attacks, etc.
There are lots of link spammers out there. Their goal is to make garbage sites seem more "important" to search engines by increasing the number of unique sites that link to them. WordPress is targeted because it is a popular enabler of web sites, that can essentially be used to make hundreds of unique sites do your bidding.
I have seen similar problems with any pervasive site technology. For instance, I've seen public wiki pages vandalized ("talk" pages with hundreds of lines of links to garbage sites, presumably to make Google find them more relevant). And I've seen problems with PHPBB forums (new "accounts" with profile links to "home pages" that are really just more spam for Google to find). These things don't happen because someone goes to one site, they happen because robots target the popular technologies and try to automatically find sites on the web that are exploitable.
You see this with closed-source pervasive technologies (like Windows). It's pretty clear that seeing source code is not necessary to repeatedly perform devastating hacks.
In all cases, the correct solution is vigilant and competent administration. Of course if you never change the defaults, you're more vulnerable. Of course if you don't patch your server regularly (these tools are constantly improved), you're open to old hacks. And ultimately, if you choose the popular tool, you can expect more frequent issues.
There are plenty of other examples of software in all kinds of languages out there that have had security holes, plenty of them were closed source too.
In particular, he notes that publishing source code makes vulnerabilities easier to find. True, obviously. I would note that it also makes them easier to fix. Now, we could get into a fight over whether "find" or "fix" is more important. But instead, let's observe that making software open-source, significantly changes the security dynamic.
Thus, we can ask, "What does it take to make closed-source software secure?" and "What does it take to make open-source software secure?" Whatever the answers may be, it is likely that they are different answers.
Essentially, the counterargument is that for popular projects (like the Linux kernel, Apache, etc.), the increased number of eyes on the code will catch more bugs, making up for the ability of attackers to read the code.
Referring to the WordPress worm in particular, Reynolds claims (second paragraph under The Problem) that it's not WordPress's fault that its software is buggy, it's Open Source's fault. That seems like claiming that it's not Pons and Fleischmann's fault that cold fusion didn't work, it's science's fault for encouraging them to publish their results. The real problem is that people run out-of-date versions of Internet-facing software.
Yes, not only is the original article wrong but it's wholly ignorant of the debate that the parent references and so doesn't advance the discussion in any meaningful way - ie, the original article is deliberate or accidental FUD.