Developer discussion for Creative Commons technology and tools

[Dan Mills <dan AT creativecommons.org>]

I'm not sure that would work, unless you are suggesting that /legalcode be the bare-bones version and we generally link to /explained instead. If so, skip to the bottom of my reply :)

The problem is that the license, when prepared for human consumption, benefits from changes in its surroundings. For example:

(1) user visits the deed. It might be required for the legal code to be included there in some form (such as an expanding widget, for example). This is what I'm hearing from the legal team - it's not 100% clear yet, but a possibility. In such a setting, the legal code would need to

- not contain the big header

- not include the "back to deed" link at the bottom

- not include links to translations being proposed for the legalcode page (the deed already has locale links)

(2) user visits the legal code page. It makes sense to have the above, plus potentially other tools for e.g. viewing errata changes in-context.

Maybe what you meant is something like:

deed: links to /explained (not /legalcode).

/explained: web readable version generated from /legalcode

/legalcode: hand-edited bare-bones document with no JS or stylesheets.

If so, that's closer to what I proposed, except that it changes what /legalcode is today somewhat. I think we should have the URLs serve the purpose they serve today, just invert how they are generated:

deed: links to /legalcode (as it does today)

/legalcode: generated from /legalcode-plain

/legalcode-plain: hand-edited bare-bones document with no JS or stylesheets.

Dan

On Thursday, April 18, 2013 at 11:30 AM, Maarten Zeinstra wrote:

I think we're almost there. 

Let's say we have one HTML document that contains the license. It is mark with css classes and anchors. Now webtechnology can be created and or exists to embed that document. What I am getting at is that something like 

http://creativecommons.org/licenses/by-nc-nd/3.0/legalcode

should consist of only one document with css designed for print, mobile devices, and regular screens, in may include js for errata

But let's only have one document! Let's not have it marked up in one markup language and presented in another, let's have one document. That's all I am saying.

Now if you want to add functionality, do not do it on the legalcode page. For example if you want to create a product say:

http://creativecommons.org/licenses/by-nc-nd/3.0/explained

You should use _javascript_ or similar tech. to fetch the legalcode page and work from that to add functionality.
 

-- 

Kennisland 

| www.kennisland.nl | t +31205756720 | m +31643053919 | @mzeinstra

On Apr 18, 2013, at 20:20 , Dan Mills <dan AT creativecommons.org> wrote:

Can you explain a bit more what you mean? Not sure I get how it's different from what I'm suggesting.

Dan

On Thursday, April 18, 2013 at 11:19 AM, Dan Mills wrote:

This is not actually about Markdown.

This is about having a minimal document that we transform into the page the user sees, by adding styling, headers, errata, links to translations, or whatever the legal team wants to add. Markdown is just one example of a format that would make that easy.

I completely agree that there should be some immutable version of the license, the discussion is about what that should look like.

Dan

On Thursday, April 18, 2013 at 11:14 AM, Maarten Zeinstra wrote:

But really that is just silly.

What legal is requesting is a way to get to the 'real' legal code, but actually they are asking for a simple representation of that code. As I see it they want something like a 'print-version' of the legal code. That would still mean that there is markup in some way.

I would have one basic html document, slightly different than Kinkades example (I would stack css classes), and have two pages. One for web viewing and another completely dressed down for print.

I think given the option of a print version or a markdown version the legal-team will probably choose the former.

-- 

Kennisland 

| www.kennisland.nl | t +31205756720 | m +31643053919 | @mzeinstra

On Apr 18, 2013, at 20:10 , Dan Mills <dan AT creativecommons.org> wrote:

No, end users would use HTML when used in a web context, of course. But consider the two use-cases I mentioned:

* Adding translation links at the bottom of the legal code page.

* Embedding the legal code into the deed in some manner.

Both are HTML for the end user, but *different HTML*.

So again, the point is that we need:

1) a *minimal* format to express the licenses themselves, and nothing more.

2) some tooling to take those licenses and format them as needed depending on context.

And although we could write that tooling today with the HTML as-is, I would *much* rather write tools that know less about the license content.

Dan

On Thursday, April 18, 2013 at 10:57 AM, Greg Grossmeier wrote:

I guess the miscommunication here is:

what tooling will you build that needs to use something other than HTML

to display the license?

What use-case do you have in mind, Dan?

Greg

<quote name="Dan Mills" date="2013-04-18" time="10:50:26 -0700">

Hi,

The licenses still contain too much information which is not actually part of the licenses. Just open that link, view source, and behold.

Of course we could parse it, but how would you decide what you can remove and what is part of the license? We need to minimize the amount of code that has such decisions embedded in it.

Dan

On Thursday, April 18, 2013 at 10:45 AM, Nathan Kinkade wrote:

All of the CC licenses validate as XHTML 1.0 Transitional. There are

a lot of really great XML parsers out there for manipulating such

documents. It would be trivial for us to clean up the HTML in the 4.0

licenses to more minimal, using better nesting of id and classes so we

can use more accurate CSS selectors, and _javascript_ can more reliably

navigate the DOM. I have already started this when I put together the

Draft 3 of the 4.0 licenses by giving a unique ID to each section and

subsection:

http://mirrors.creativecommons.org/drafts/by-sa_4.0d3.html#s3a1B

Nathan

On Thu, Apr 18, 2013 at 1:30 PM, Dan Mills <dan AT creativecommons.org (mailto:dan AT creativecommons.org)> wrote:

Hi Bjorn & Maarten,

I think you're missing a key point that Kat is making: the legal team is

looking to change the pages that the licenses are on, to add translation

links. I also know that they are thinking about ways of including the

licenses inside the deeds, which would also require some changes to the

license pages.

So the point is not "what do you think of Markdown" in a vacuum, it's: what

format can we store that contains only the absolute minimum to be

considered to be part of the licenses, so that we can build tooling to style

it appropriately depending on the context.

We could obviously write tooling that takes the current HTML and transforms

it, but such tooling would need to be highly content-aware: it would need to

know which parts of the HTML file it can remove or change, and which ones it

cannot. We likely can't eliminate that completely regardless of the format,

but we should try to minimize it.

Markdown seems pretty close to the minimal format that lets us express what

we need. We could also continue to use HTML, but we'd need to use a minimal

subset--not what we use now (which includes images, scripts, links not part

of the license, etc).

So, looking at your (Maarten's) four points with this in mind:

1. Both markdown and HTML (HyperText Markup Language) are markup

languages, it seems silly to convert one markup language into another.

This is not a criteria for choosing a format to use.

2. Adding markdown to the infrastructure creates extra dependancies on

a conversion between markdown and HTML, one that will probably takes

more skill and time than doing these licenses immediately in html

It does, but the alternative is an HTML->HTML transformation, which is

arguably more complex because HTML is so expressive. We could make it work

if we enforced a very limited subset of HTML as the input, though.

3. Markdown is not a standard and we cannot rely on it to stay the

same, HTML is.

This is not actually true, the history of HTML is littered with examples:

http://en.wikipedia.org/wiki/Comparison_of_layout_engines_%28Non-standard_HTML%29#Deprecated_HTML_elements

But you're right that Markdown is not currently led by any large standards

body. I think there are two mitigating factors:

(a) The primary uses for these files will be:

- to be transformed for general consumption, and

- to serve as an archive.

The first is internal to CC only, the second requires at most that the file

be readable at some point in the future without our help. In other words, we

do not need every client (browser) to natively understand the format.

(b) Markdown is so incredibly simple, it's hard to imagine a future where

someone will be unable to read it:

http://etherpad.creativecommons.org/p/markdown-example

4. Markdown basically is short hand for HTML, again why would we use it?

Simplicity, and as a forcing function to get us to stop putting in

extraneous content in our licenses.

Dan

_______________________________________________

cc-devel mailing list

cc-devel AT lists.ibiblio.org (mailto:cc-devel AT lists.ibiblio.org)

http://lists.ibiblio.org/mailman/listinfo/cc-devel

_______________________________________________

cc-devel mailing list

cc-devel AT lists.ibiblio.org

http://lists.ibiblio.org/mailman/listinfo/cc-devel

--

| Greg Grossmeier GPG: B2FA 27B1 F7EB D327 6B8E |

| http://grossmeier.net A18D 1138 8E47 FAC8 1C7D |

_______________________________________________

cc-devel mailing list

cc-devel AT lists.ibiblio.org

http://lists.ibiblio.org/mailman/listinfo/cc-devel

_______________________________________________
cc-devel mailing list
cc-devel AT lists.ibiblio.org
http://lists.ibiblio.org/mailman/listinfo/cc-devel