RadHub beta

Dear all,

as promised, I am pleased to announce RadHub, my proposal for a
referencing and indexing system of RADIANCE resources that are not hosted
centrally.

--> http://radhub.luxal.eu <--

Please try it out by submitting/voting/sifting through. During the Beta
testing stage, please ignore the requirement of having the RadHub logo
with uplink on your pages (Submit, Step 2). This check is disabled until
we go live, hopefully at the workshop. The data base will be reset then,
and we'll start from scratch. So don't be shy of submitting any pages,
even if it is not directly RADIANCE related. The more, the merrier.

Please send lots of feedback.

Axel

Hi Axel,

Please send lots of feedback.

I particularly liked the "Made in Portugal" icon, but you should
link it to somewhere relevant :wink:

First troubles, I tried with my site, changed the metatags (it's tricky,
so it might not make your idea very popular) but it didn't work:

RadHub Submit Confirmation
Action : submit
rawSite -> http://www.bozzograo.net/radiance/

Error: Strange URL: http://www.bozzograo.net/radiance/

Now I know I have a strange url, that's interesting ...

Obrigado,

Francesco

Hi ya,

I particularly liked the "Made in Portugal" icon, but you should
link it to somewhere relevant :wink:

Yes, there are not many things Made in Portugal, other than sunburns on
pale English (and German) skin... But what/where is 'somewhere relevant'?

First troubles, I tried with my site, changed the metatags (it's tricky,
so it might not make your idea very popular) but it didn't work:

I'll look into it as soon as the site is back up. I hope it wasn't us
crashing it. The hosting seemed liked a good deal, but they are just not
up to anything.

A sua saude

Axel

Right,

the hosting has got itself together againg, and the answer is......

rawSite -> http://www.bozzograo.net/radiance/

Error: Strange URL: http://www.bozzograo.net/radiance/

The FULL URL is required, including 'index.html' or (in your case) index.php.

I've added your site and also a note to the submit box. You might want to
boost the rating now by voting 'Excellent' several hundred times.

First bug is squashed! Thanks for your feedback.

Axel

Axel Jacobs on 27 July

Yes, there are not many things Made in Portugal, other than
sunburns on pale English (and German) skin... But what/where
is 'somewhere relevant'?

this for example:

http://www.ivp.pt/index.asp?idioma=1&

And of course Google knows a lot more ...

> First troubles, I tried with my site, changed the metatags (it's
> tricky, so it might not make your idea very popular) but it
didn't work:

Indeed I looked into the procedure and stopped there because
it will require some rewriting of source code (or in my case
of a html generator) which could be quite a work for larger
sites with pages for different topics.

I understand the wish to allow only authorised people to submit
new pages but wouldn't it be a bonus if others could submit
interesting pages they found on there web searching tours?
Perhaps these submits should be approved before they are allowed
into the database. I expect, though, that for your scripts to
work - and in fact the whole logic behind it - the presence of
the META tags is required.

A sua saude

Ah, your portuguese is improving ...

Thomas

···

***********************************************************************************
This e-mail, (and any attachments) is confidential and may be privileged. It may be read, copied and used by the intended addressee only. If you have received this in error please contact BDP immediately.

If you have any queries, please contact the sender.
***********************************************************************************
Building Design Partnership
Registered in England No 2207415:
Registered Office: Building Design Partnership Ltd, Sunlight House, PO Box 85, Quay Street, Manchester, M60 3JA, http://www.bdp.co.uk
***********************************************************************************

> First troubles, I tried with my site, changed the metatags (it's
> tricky, so it might not make your idea very popular) but it
didn't work:

Indeed I looked into the procedure and stopped there because
it will require some rewriting of source code (or in my case
of a html generator) which could be quite a work for larger
sites with pages for different topics.

How about if I alternatively scanned the page for something like
<p id="keywords">....</p> and
<p id="description">....</p>, in case there are no valid meta tags.

You could then hide the two paragraphs with your CSS.

I understand the wish to allow only authorised people to submit
new pages but wouldn't it be a bonus if others could submit
interesting pages they found on there web searching tours?
Perhaps these submits should be approved before they are allowed
into the database. I expect, though, that for your scripts to
work - and in fact the whole logic behind it - the presence of
the META tags is required.

Well, yes and no is the answer. RadHub is meant to be
a) fully automated
b) indexable by keywords
c) ordered by popularity
d) community-driven

So somebody will have to sit down and think about the keywords. And who
would be better suited than the author of the submitting site?

With regards to submitting a 3rd party site, what I had in mind is that
people would get in touch with the author, and e-mail them a copy of The
Missing Manual cover page (See under Art Work), asking them nicely to have
their pages indexed.

Ours is still a fairly small community, and many people know one another
either personally or from the e-mails they exchange.

Axel

Um, that IS a "full" URL. URL's can be almost anything. They don't
necessarily have to point to files and you shouldn't assume they do. Why
does your web site have such a silly requirement on URL's anyway?

···

On Thu, 27 Jul 2006 01:27:00 +0100 (BST) "Axel Jacobs" <[email protected]> wrote:

Right,

the hosting has got itself together againg, and the answer is......

> rawSite -> http://www.bozzograo.net/radiance/
>
> Error: Strange URL: http://www.bozzograo.net/radiance/

The FULL URL is required, including 'index.html' or (in your case)
index.php.

Evening all.

Let the evening mail session begin ...

How about if I alternatively scanned the page for something like
<p id="keywords">....</p> and
<p id="description">....</p>, in case there are no valid meta tags.

You could then hide the two paragraphs with your CSS.

No changes for me, please! I already had META tag support
in the script, I just dropped it during the last rewrite.
Adding it in again is probably only a matter of the right
printf statement.

The problem I see is that some activity from the side of
the site admin is required. This will stop most good content
out there from being added to the database due to laziness
or lack of time.

I understand the wish to allow only authorised people to submit
new pages but wouldn't it be a bonus if others could submit
interesting pages they found on there web searching tours?
Perhaps these submits should be approved before they are allowed
into the database. I expect, though, that for your scripts to
work - and in fact the whole logic behind it - the presence of
the META tags is required.

Well, yes and no is the answer. RadHub is meant to be
a) fully automated

... except for the fact that someone has to nag someone else
to do two other things to get the site registered. Google
works without that.

b) indexable by keywords

Google does that.

c) ordered by popularity

Google does that, too.

d) community-driven

Ah! Don't ask what RadHub can do for you ...

Sorry, I may read a bit sharp here. I think it's a good
thing what your trying to do. But this won't stop me thinking
about how it could be made better which - for a community site
- usually means easier to use.

So somebody will have to sit down and think about the keywords.
And who would be better suited than the author of the submitting
site?

Agreed. But what if the author can't be bothered with the idea
or doesn't have any influence on the way the document is published?
A good RadHub will not remain restricted to pure Radiance. Think
about special HDR cameras or photometers. Companies selling these
products might outsource their internet presence or even use the
service of a web designer (shudder)!

With regards to submitting a 3rd party site, what I had in mind is that
people would get in touch with the author, and e-mail them a copy of The
Missing Manual cover page (See under Art Work), asking them nicely to have
their pages indexed.

Nice cover. Did remind me of the Radiance book I always wanted
someone else to write :wink:

This submitting stuff could be simplified (for the submitter)
with another form which takes the URL and (ideally) the email
address of the webmaster and sends a polite invite along the
lines of:

"Dear webmaster of shining.site.com

Your site has been recommended for addition to the RadHub
online database (see http://radhub.luxal.eu). There is nothing
to be gained but fame, there are no costs but a few minutes of
your time.

If you would like to add your site please follow the procedure
outlined on http://radhub.luxal.eu/submit.shtml.

Thank you for your time.

Your RadHub bot"

Combined with some spam prevention and tracking (we don't want
to spam people who obviously don't have an interest in adding
their sites, do we?) this would make it a one copy-paste-click
action to submit a good site.

Ideally this should be enough for the RadHub bot to jump into
action and get some sense out of the submitted URL (using meta
tags or the full text content). You can add admin approval after
that if you wish but, in essence, that's what the community rating
is for.

Ah well, these are my thoughts about the topic which are not
only RadHub thought but have evolved thinking about the structure
of our department's image library - which is good as long as you
know in advance what you're looking for and where to find it ...

Thomas

···

On 27 Jul 2006, at 13:17, Axel Jacobs wrote:

Dear all,

Um, that IS a "full" URL. URL's can be almost anything. They don't
necessarily have to point to files and you shouldn't assume they do. Why
does your web site have such a silly requirement on URL's anyway?

I beg to differ: a URL is the location of a FILE on the Internet. The
reason why we don't normally have to worry about the last part, i.e. the
actual file name, is that the web server is configured to serve a default
page, which is usually 'index.html', 'index.shtml', or 'index.php', if no
page is explicitly given.

If you remember, I asked for your feedback, and what you all have said has
shown me that what I've presented is obviously not what you all are
looking for.

Talking about feedback: Ian, next time, you might want to consider making
sensible SUGGESTIONS, rather than saying that something is SILLY, when you
don't even know what you are talking about. Is it really so difficult to
say that the submit form SHOULD gracefully handle choped-off URLs?

Thanks again for all the feedback. I am pulling the plug on RadHub. I hope
I haven't wasted too much of your valuable time.

Long live Google (which has some good definitions on 'URL'...)

Axel

Well I beg to differ as well. The URL spec written by Tim Berners-Lee back
in 1994 says this:

   This section describes the syntax for "Uniform Resource Locators"
   (URLs): that is, basically physical addresses of objects which are
   retrievable using protocols already deployed on the net.

Objects, not files. Specifically about HTTP URLs:

   The HTTP protocol specifies that the path is handled transparently
   by those who handle URLs, except for the servers which de-reference
   them. The path is passed by the client to the server with any
   request, but is not otherwise understood by the client.

Here's the full spec:
http://www.w3.org/Addressing/URL/url-spec.txt

It's important that someone creating a website, especially with dynamic
content, understands how HTTP works. And you almost had it there when you
said "the web server is configured to serve a default page". Yes, it's the
web server that decides what to do with the URL! A lot of URL's map to
files on the web server, but a lot don't.

How does e.g http://en.wikipedia.org/wiki/Uniform_Resource_Locator
map to a file? All of Wikipedia is in a huge database. It's the
server-side software that decides what to do with each URL, not the client.

Sorry to hear that you're pulling the plug on RadHub so soon. I hope my
technical nit-picking wasn't a major part of that decision. I just have to
point out errors where I see them. I didn't think it was that big a deal.

bye

···

On Thu, 27 Jul 2006 23:43:01 +0100 (BST) "Axel Jacobs" <[email protected]> wrote:

Dear all,

> Um, that IS a "full" URL. URL's can be almost anything. They don't
> necessarily have to point to files and you shouldn't assume they do.
> Why does your web site have such a silly requirement on URL's anyway?

I beg to differ: a URL is the location of a FILE on the Internet. The
reason why we don't normally have to worry about the last part, i.e. the
actual file name, is that
page, which is usually 'index.html', 'index.shtml', or 'index.php', if no
page is explicitly given.