Re: 1.4 f: terms for links colocated with their ends
Liora Alschuler wrote:
> At 11:49 AM 2/7/97 +0700, James Clark wrote:
> >At 11:09 31/01/97 -0800, Tim Bray wrote:
> >>1.4.f Should we define terms for links that are colocated with their
> >> ends,
> >> and if so, should we use in-line and out-of-line?
> >No. I find "out-of-line" much less suggestive of the nature of the link
> >that the existing HyTime term "independent".
> >I think using the term "in-line" is going to be confusing when we get to the
> >style stage, since the term "in-line" is used very heavily in formatting (at
> >least in DSSSL). I am not very keen on the term "contextual" either. Since
> >the "in-line" links are the kind of links that most people are currently
> >used to, may be they could be called something like "simple", "normal" or
> I also had much trouble with these terms.
> I believe that what is called here "in-line" and in HyTime "contextual" are
> links that share a number of characteristics:
> 1. located at the (start) link end
> 2. uni-directional
> 3. hardwired
> Choosing 1. -- co-location -- as the basis for naming is confusing for the
> reason James cites and also obscures the other characteristics which
> identify this type of link. I can't think of a single name that will evoke
> all three characteristics and I agree that "simple", "normal" or "basic",
> which require definition, are an improvement on "in-line".
> "Out-of-line" sounds like scolding and emphasizes the location of the link
> look-up table, which is much less relevant than the fact that there is a
> look-up of some sort.
There can be a lookup in either case.
> "Simple" and "independent" are somewhat descriptive and are explainable.
Yes. Basic may be better. Simple makes people think it isn't powerful.
> Is there a reason why we can't just call these "direct" and "indirect" links?
Yes. A clink can point to a location ladder. Direct and indirect are
not actually descriptive of functionality. Inline and out-of-line
are just descriptive of storage location. Neither has much descriptive
power. Context and independent don't have very much, but they are in
current use and used correctly. They are also standard. The XML
specification is just a set of recommendations that COULD be
It will be simpler to achieve this later process if standard terms
are used. Durand argues against 'de novo' terms, but for a pre-standard
practice, all terms are de novo. Better to use the standard.
> This also keys on just one of the three characteristics of this type of
> link, but if
> 1. a direct link is always located at its start link end
> 2. a direct link is always uni-directional
> 3. a direct link is by definition hardwired
Define "hardwired". Is the <a href="somefunction here"> being generated
by a server side script "hardwired"?
> then, I think we should call it a direct link because unlike co-lation or
> uni-directionality, it does imply the other two characteristics.
> There may be some very obvious reasons that "direct" and "indirect" are
> non-starters, but they sure would speak well to both technical and
> non-technical audiences.
Call them George if you like. The average user tries them once, and
uses that "experience" to identify functionality. Stick with the
terms already in use in the larger SGML community. What is being
suggested now is guaranteed to perpetuate confusion and even more