Tomorrow, Oregon House Bill 2892 (also here) goes before the General Government Committee for discussion. If successful, it will then go before the Ways and Means Committee, then to the House floor for a vote.
I'm particularly interested in the discussion of open standards and document formats. Open Source and Free Software have several advantages here. The text of the bill mentions a few. Here's what I see:
Anything I've overlooked?
Re:Security
chromatic on 2003-04-02T20:13:46
That's a very good point. I'm certainly all for transparency in government and government software. It's also mentioned in the bill. Other people will likely cover that in their testimony and supporting documentation.
Is there a way to cover the auditability aspect of Open Source while discussing open standards and protocols? I'd like to stay within that narrow topic -- it helps to be laser-precise when talking to lawyers and legislators.
:) Re:Security
ziggy on 2003-04-02T20:51:10
Perhaps. Is there any chance you can get Whitfield Diffie on your side? (On second thought, he's a pretty great mind and a deep thinker, but he rambles almost incomprehensibly at times.)Is there a way to cover the auditability aspect of Open Source while discussing open standards and protocols? I'd like to stay within that narrow topic -- it helps to be laser-precise when talking to lawyers and legislators.:) Diffie's argument revolves around a fundemental tenet of security: a secret is only as good as your ability to change it. For example, suppose you have a combination lock on the front door of the office. You may consider it secure because it cannot be cracked(!), and furthermore it is a closed, proprietary combination lock -- no one knows what goes on inside.
Now suppose you have 100 employees in your facility. Each and every one of them knows the combination. That's fine, because each and every one of them is trustworthy(!) and no one would squeal.
Now suppose your impenetrable combination lock gets cracked. What do you do? You can't change the combination because your 100 employees won't be able to get into the office anymore. Because you can't change your cracked combination, someone else who is not trustworthy can enter your building any time he wants without your permission, simply because you cannot change the combination.
Oh, and let's not forget that an untrustworthy cracker is unlikely to be afraid of breaking the law or even the Patriot act....
This analogy is the basis of Diffie's refutation of security by obscurity. A closed source package has its source code obscured. You cannot change it easily. It will have security holes -- every piece of software of value does. However, you cannot fix them -- only your vendor can fix them, and only on his timescale. Furthermore, because closed source software relies on a secret that cannot be easily changed (the binaries), it is fundementally insecure.
Open source, on the other hand, does not have this property. It is a secret that can be easily changed (much like a PGP key can be revoked and reissued). It is open to inspection, so it is more likely people will find security bugs. Furthermore, because it is easily changed, it is more likely to keep your secrets.
(I hope that makes some sense. It was early in the morning and Diffie was wandering pretty far afield when I heard him speak...)
Re:Security
brev on 2003-04-03T04:35:50
Changeability: that is a brilliant insight.
Excellent refutation of biometrics for free, too.
Unfortunately security-by-obscurity seems to be the first instinct of every layperson out there.
Perhaps it's because their only experience of security is dealing with passwords. They know that they are supposed to be all tricksy with their passwords, so they have the sense that computer security == being tricksy.
Perhaps one should meet that head on, explain that the ideal security system has no secrets at all, except for the password.Re:Security
wickline on 2003-04-02T21:21:16
> cover the auditability aspect of Open Source
> while discussing open standards and protocols
HTTP Basic Auth clearly defines how your auth info is sent over the network, and so folks know not to use it for Important Things. If it wasn't defined, you might just cross your fingers and hope it was sufficiently secure. Instead, you know to use digest auth or basic auth over HTTPS.
A DTD can clearly define the content of a document, and that document can be determined to be well-formed by freely available tools. If the DTD is sufficiently strict, you can be sure that no mystery data is being included accidentally. Compare this with the fuss some folks made a number of years back when they learned that MS Word documents contained
- the path to the document on their machine
- information about which printers they used
- their contact information
- occasional random hard drive data from allocated
storage space which hadn't been zero'ed out
-mattRe:Security
vsergu on 2003-04-02T21:30:42
Another worrisome thing MS Word documents can contain internally is text from previous versions that the user thinks has been completely deleted, especially if "Fast Save" is enabled (the default configuration, at least at one time).
There's a similar example with the IRS -- all forms are made available in PDF.
There's also a growing body of experience in Europe using Open Source.
That's about all I can think of at the moment about open document formats. There's a lot more to be said about using open standards and open source.
Re:The role of Government
gnat on 2003-04-03T11:24:39
You're thinking of Tim Bray. See "XML Confers Longevity" in this blog entry.I vague remember someone reminiscing a question that came up at some sort of government and technology event recently--Nat
Re:The role of Government
jdavidb on 2003-04-03T18:30:26
I won't mention that the world's most preferred SGML tools are open source
And the nice thing is that you don't have to. Anyone getting into those will find them anyway.
Sometimes I think if the open source community just keeps quietly creating better and better stuff, they will eventually win the war.