Re: [Terminology] [Gendispatch] WG Review: Effective Terminology in IETF Documents (term)

[Eliot Lear <[e-mail]>]

[I am reducing this from gendispatch in accordance with the IESG’s request.]

> On 25 Apr 2021, at 01:48, [e-mail] wrote:
> 
> (I am a new subscriber, so I apologize in advance if I break any conventions.  Thanks.)

Welcome to the party, Tom.

> 
> A couple of points concerning a list, I would think that any list of good words would have to say the bad words with which the good correlate.  Otherwise one is in the position of a person trying to guess from contextuual clues what a redacted word might be.  Which is common in living languages, but is difficult without some amount of shared context.  But including the bad words leads to the problem of auto-filtering eliminating the list; which leads to the need for a special status by which bad words can be identified as being merely written about.  (This is the usual problem of confusing meaningful use of a word versus merely talking about the word as a word, ie mentioning it.  There is a lot of philosophy and logic about this.  The usual solution is to distinguish mention cases by identifying with some or other quote mark--including finger quotes--any occurrence of a word being mentioned.)


> 
> As for having a list, I think any occurrence can be taken to be a list, eg a singleton word-pair can be taken to be a one-member list, normal in computing.  If the example is in regular prose writing, there is more context than there is for the usual idea of the efficiency of the list form.  A hybrid could be a list entry being a word-pair plus a brief example usage.

Mostly what you are raising above is about auto-substitution.  I have seen some tools that do just that sort of thing, and in fact, the place where I really want the help as an individual is in the same logic that I can find with my spell checker.  [Such an addition would be particularly useful for pronouns, especially if I have indications in my address book of what a person’s preferred pronouns are.]

I think you’re also coming at this from the right angle, in that, given a list, tooling would at least flag the bad words, just as our model checkers flag bad yang, and our Internet Drafts nits checker flags common errors.  However, as with those other tools, I suggest that the “fixing” part should be left to the authors and WGs.

> 
> As for controversy, these are controversial times.  If IETF wants some internal advice people--internal or external--can refer to, why not try to do one's part to help.

Within the limits of one’s expertise, I would agree.  As we are not subject matter experts at establishing principles for this sort of thing– our community is comprised on the whole engineers, “external” seems to me to be the wise choice, so that we can benefit from those who are experts.  If by some strange stretch we were to attract a reasonable pool of such experts, I would of course feel somewhat differently.  Even then, we would have to guard against we engineers interjecting our layman views.

Eliot