[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [Ltru] Re: draft-ietf-ltru-matching-02



Jovial in-line response below.

Addison

Addison P. Phillips
Globalization Architect, Quest Software
Chair, W3C Internationalization Core Working Group

Internationalization is not a feature.
It is an architecture. 

> -----Original Message-----
> From: ltru-bounces at lists.ietf.org [mailto:ltru-bounces at lists.ietf.org] On
> Behalf Of Frank Ellermann
> Sent: 2005?6?15? 7:27
> To: ltru at ietf.org
> Subject: [Ltru] Re: draft-ietf-ltru-matching-02
> 
> Addison Phillips wrote:
> 
> > Some corrections entered as shown below. New copy online.
> 
> Okay, but Mark already convinced me that I got Randy wrong in
> my premature "last call" frenzy.
[Addison Phillips] 

Yes, but good edits are useful to capture.
> 
> >> - the credits are dubious
> [...]
> > Ahem... and both co-editors.
> 
> AFAIK editors do not credit themselves, it's obvious that they
> contributed in some way.  But I'm currently at odds with some
> of the professional standards of the I"E"SG and studying texts
> like RfC 2026 6.5.2 and RfC 3777 chapter 7.
[Addison Phillips] 

No, they don't usually do so. But in making a list, one should be thorough.
> 
>  [security considerations]
> > It is NOT a simple copy: language ranges
> [...]
> > are the item addressed by the security section.
> 
> Almost identical parts (matching -02 vs. registry -04):[Addison Phillips] 
[Addison Phillips] 

Yes, nearly identical, but that isn't a reason not to have the text.
> 
> Where is the protection of the IANA tag registry against DoS
> attacks relevant for the purposes of "tag matching" ?
[Addison Phillips] 

It isn't applicable to matching. The registry is never even mentioned in the matching draft, IIRC.
> 
> | This is a special case of the general problem that anything
> | you send is visible to the receiving party.  It is useful
> | to be aware that such concerns can exist in some cases.
> |
> | The evaluation of the exact magnitude of the threat, and
> | any possible countermeasures, is left to each application
> | protocol.
> 
> And how does this almost literal copy from the registry draft
> affect any tag "matching" applications ?  It's somehow like
> saying "regular expressions might be abused to guess what the
> user actually wants".  Now, yes, that's their purpose, and the
> privacy issue is addressed in the registry draft.
[Addison Phillips] 

The privacy issue really pertains to tracking the requests people make and those are expressed as ranges. In other words, language tags by themselves do not have privacy issues. If I tag some XML content with xml:lang, that has no privacy issue. Neither is labeling this email with a Content-Language: my "identity" (in the form of my email address) is a much better indicator of who I am. It is a privacy issue when you request things with a token that may be somewhat unique and thus can be tracked. The token is a range, not a tag, and thus matching applications may very well also be tracking applications.
> 
> IMHO all you need is the first paragraph with a pointer to the
> registry draft:
> 
> | Language ranges used in content negotiation, like any other
> | information exchanged on the Internet, might be a source of
> | concern because they might be used to infer the nationality
> | of the sender, and thus identify potential targets for
> | surveillance.  This issue is also addressed in [RFC3066bis].
> 
> No fancy stuff about the lack of "security considerations" in
> 1766 and 3066, just the facts, ready.  Especially don't claim
> that this is the _only_ known "security consideration", there
> are already more in the registry draft.
[Addison Phillips] 

That is good text.
> 
> > Better in matching (where ranges are defined) than in the
> > registry draft!!
> 
> Yes, above I took the registry wording with an s/tags/ranges/.
[Addison Phillips] 

Yep, I'm still putting in the formulation "tags and ranges" (since matching also concerns tags---it is matching tags to ranges).
> 
>  [recommended order of sections in draft 2223bis -08]
> > Where should I put it?!? It isn't clear from the cited draft,
> > I dislike the endless arcana.
> 
> Me too, I hope that they will settle on one clear document, at
> the moment there are at least three (draft-hoffman-rfc-editor-
> author-guide, draft-rfc-editor-rfc-2223bis, 1id-guidelines).
> 
> And if you think that you might be up to speed Bruce will find
> some relevant errata, some of them are acknowledged but still
> unpublished.
> 
> The "recommended order" in 2223bis-08 (expired February 2005):
> 
>     7a.  Contributors              7b.  Acknowledgments
>     7c.  Security Considerations   7d.  IANA Considerations
>     7e.  Appendixes                7f.  References
[Addison Phillips] 

And there was much rejoicing. Not only will I have to reorder this draft, but also the registry draft.
> 
> 7c is required.  7d is AFAIK also required, but not in this
> draft.  BTW, they just discuss this oddity on the IETF list.
[Addison Phillips] 

I saw the thread on null IANA sections.
> 
>   [no references in the abstract]
> >> I'm not joking, I discussed this on the xml2rfc list.
> 
> > GROAN. I really REALLY dislike endless arcana like this.
> > The abstract for the registry draft cites RFC 3066 (duh).
> 
> That's a matter of the DTD, it doesn't allow <xref> in the
> <abstract>.  It's also a rule in the expired 2223bis-08 4.5:
[Addison Phillips] 

Both documents cite RFC 3066 (and registry cites RFC 1766), but not using xref.

By the way, the DTD *does* allow xref in the abstract. I'm using XMetal Author, so there is no way for me to create invalid XML.
> 
> | An Abstract should be complete in itself; it should not
> | contain citations unless they are completely defined within
> | the Abstract.
> 
> >> e.g. <http://www.inter-locale.com/>
> [...]
> > We both hate spam. I exposed my email address and that's
> > sufficient.
> 
> We all do, but some of us love googlebot.  Do you want me to
> delete my link to your page, only because an address harvester
> might find and follow it ?  I meant the optional <uri>..</uri>
> part, not your munged <email>..</email> address, that's fine.
[Addison Phillips] 

No, the website isn't really a spam generator. But we have been conservative in what we placed in our personal profiles. The Inter-Locale address is useful in this context because it hosts various editing copies.
> 
>   [unused 2119 keywords]
> >> Bruce checks this.
> > Argh. Let him.
> 
> Okay.  And if Scott has a problem with the lower case variant
> as soon as the upper case is defined let them shoot it out ;-)
[Addison Phillips] 

I did go through and do the fixes for that. I left the full list since we're not near last call yet. We may have need of various OPTIONAL keywords :-).
> 
>  [maybe replace ( "x" / "X" ) by "x"]
> > No, it shouldn't, I don't believe.
> 
> Maybe.  But never ever try to say "x" if you don't want also
> "X", they'd bury you under tons of RfCs and "hum" to it.  And
> draft taobis-02 is *not* expired, no excuses, no prisoners :-)
> 
I was keelhauled and burnt at the stake in a previous draft over this silly issue, so I'm slightly loathe to remove the case distinctions. Besides, look at the registry draft.


_______________________________________________
Ltru mailing list
Ltru at lists.ietf.org
https://www1.ietf.org/mailman/listinfo/ltru




Note Well: Messages sent to this mailing list are the opinions of the senders and do not imply endorsement by the IETF.