[Top][All Lists]

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

Re: lynx-dev hidden links

From: Klaus Weide
Subject: Re: lynx-dev hidden links
Date: Mon, 19 Jun 2000 14:36:46 -0500 (CDT)

On Mon, 19 Jun 2000, Eduardo Chappa L. wrote:
> [...] You still don't get it. One site is an example, and I
> have given example sof many sites that do use them for reasons that are
> unknown to me, and I hope to you too.

You _hope_ that I don't know more than you?  That's not very
constructive. :)

No, I don't _know_ why any given hyperlink is coded in a way that
makes it a "hidden link" to lynx.  But I can make assumptions, based
on the specifics of each case.  So far, I have come to the conclusion
that nearly all of them are meant to be hidden, or explicitly coded
in a way that they won't show up in a non-graphical browser.

> *** Klaus Weide (address@hidden) wrote in the Lynx list on Jun 16, 2000:
> :) In the CNN site they don't result from images without alt attributes.
> :) They result from images with explicit alt="" attributes.
> Pardon me? let's see, as of this morning I see this link:
> and the text linking to it is:
> <a href="/cnnsi/golf/2000/us_open/index.html">
> <img src="/images/0006/cnnfront_plus_tiger.jpg"  
> WIDTH=55 HEIGHT=45 ALT="" BORDER="0" align=right
> hspace="2" vspace="2"></a><DIV class="pad">
> <a href="/cnnsi/golf/2000/us_open/index.html">
> Read about Tiger Woods' U.S. Open tour de force</a>
> I see an ALT="" very clear there, I hope this is more than enough of an
> example.

Just what I said: not an image *without* alt attribute, but an
image with an explicit ALT="" attribute.

> Please don't argue back that there is a link text link to it,
> that's not the point.

That *is* part of my point.  The hidden link is just a redundant
repetition, in this case.  As in other examples that I have looked at.

> The point is that if there is a hidden link,
> information *may* be lost.

Rendering a HTML document means loss of information.  If you want to
see all the information in an HTML document, you have to look at the
source.  (Think of all the JavaScript, for example, that is getting
"lost" when lynx renders.  Or attributes that don't cause lynx to
"do" anything.  Or just comments.)

The more important question is not whether information is lost, but
whether important or significant information is lost.  What's
significant and what isn't depends on the reader, the reader's
situation, and context.  Apparently you think that "hidden links"
should always be presumed to be significant.  I disagree, and you
haven't given any examples to convince me otherwise.  (Not that I've
looked at all the URLs you gave - I have examined all the hidden
links from one document, which is more than you did.)

> [snip]
> :) So, 9 "hidden" links, in addition to 252 regular ones.  Looking closer,
> :) 5 are very dubious-looking URLs that _I_ normally would never follow
> :) ('Type=click&ProfileID=1018...&Redirect=...', no thanks), certainly not
> :) without a clear indication in the context that tells me what it's good
> :) for (and which is of course missing - we're talking about hidden links,
> :) after all).  The others (and at least one of those 5) are repeated as
> :) normal links.
> So you are saying:
>  - Out of "n" links "m" are links that *I* would never follow, therefore
> *NOBODY* should/has to follow, therefore it is not necessary to warn
> anyone about it. Pardon me? are you Bill Gates, what's good for you is
> good for the rest of the world?

I am not preventing anybody from following anything.  In fact, I have
made contributions that make it somewhat easier to follow hidden links,
should you choose to do so.

But yes, I think it is "not necessary" to warn the normal lynx user[*]
about the presence of hidden links.  I even think that such warnings
would be more annoying than useful for most users.

Maybe it is necessary to warn *you* about the presence of hidden links.
Fine - then do something about it, or get someone else to do it, don't
just complain when someone disagrees.

To answer your questions: Yes, I pardon you.  No, I am not Bill Gates.
It is more likely that what's good for me is good for a majority of
others, too, than that what's bad for me is good for a majority of
others, _in this case_, since I don't consider my reqirements and
reasons for using lynx very unusual, _in this respect_.

[*] Do you find it condescending if I talk about a "normal user", and
have some ideas about "normal" use of lynx?  If yes, consider the
alternatives.  I could make no assumptions about anything and anyone,
and consider only my own requirements (I doubt you'd prefer that).
Or poll all lynx users to determine what they want, and implement
everything they want (impossible, and why should anybody want to).

> :) So you already have 252 links to follow, and you still think that's
> :) not enough...  Aren't you happy lynx is hiding at least some of the
> :) clutter?  I am, usually.
> :) 
> The problem is not how many links a  web page has, that's pretty
> condescending, you should worry that all links should be accessible in an
> easy way,

I think I've done my part, by preserving a mode in which you can still
detect their locations and access them by numbers (-hiddenlinks=merge).

> what you browse is just one example of what a person can do, not
> what everyone should do.

My opinons are based on my experiences.  Your opinions are based on your
experiences.  That includes, for both of us, opinions about how lynx
should behave, what would be useful to have in lynx, etc.

> :) On Fri, 16 Jun 2000, Eduardo Chappa L. wrote:
> :) > I do not see why you insist
> :) > that this is not important, the bottom line is that lynx is losing part 
> of
> :) > its usefulness not reporting that there are hidden links, some of them
> :) > we may assume were intended to be seen. 
> :) 
> :) You may assume, but in my experience the assumption usually turns out to
> :) be wrong.  (sole exception: the pages given by AF)
> Is't not an assumption, it's a fact. 

Is it a fact that "some of [the hidden links] were intended to be seen"?
In a non-graphical user agent?  Or were they perhaps _intended_ to be hidden?

Where are your examples?

> :) >   What I am suggesting is that lynx gives somehow an information which
> :) > turns out to be useful, any reason why not?
> :) 
> :) Because nobody who would potentially do it has found it important
> :) enough?
> Maybe nobody has realized its importance...

You have told us that it is important for you.  Maybe someone will agree,
I haven't seen that so far.

> :) Because adding some cryptic indicator (like a character on the top line)
> :) wouldn't achieve what you want 
> Sure it would!, can you explain why not?

You wrote:

|   Just to mention an example of this kind of behavior, take a look at
| pine4.21, it has a feature today that warns you about headers in a message
| that are related to list management (how to request help, subscribe,
| unsubscribe, archives, etc). If pine did not warn you, you would never
| look for this information in the headers (OK, maybe you would, but this
| is not what the average person would do).

I haven't seen how pine does this, but I assume the information is
explicit, not cryptic.  A cryptic indicator would not convey anything
to the "average person".  It would likely only be meaningful to the
kind of person who would already "look for this information" anyway.

> :) a more explicit message would be (a) more trouble to add,
> Computers figure out that so fast you wouldn't even notice it.

Computer programs still have to be written by humans.

> :) and would (b)
> :) just clutter the interface and would "punish" most users who have no
> :) interest in seeing it?
> Then make it configurable, a command line option or whatever you think
> appropiate.

I don't object to somebody doing that.   I've just been telling you
why I won't do it, and why I don't think that it is generally useful.
You disagree.  That's fine, go ahead and make patches, start discussing
the "how to" if necessary, submit patches.


; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to address@hidden

reply via email to

[Prev in Thread] Current Thread [Next in Thread]