[Haskell-cafe] [ANNOUNCE] Fmark markup language

[José Lopes]

Hello everyone,

I just wanted to share a package I created called Fmark, now available
on HackageDB.

Feedback both on the project and on the code is greatly appreciated :)

Fmark (Friendly Markup) is a very simple markup language without
syntax and simple but sophisticated document styling, capable of
producing PDF and XML files.

The key philosophy behind this markup language is to eliminate the
strange syntactic characters seen in most markup languages, but
at the same time try to maintain a high level of expressiveness, using
only document reconstruction.

Check it out
http://hackage.haskell.org/package/fmark
https://github.com/jabolopes/fmark

Best regards,
José

--
José António Branquinho de Oliveira Lopes
58612 - MEIC-A
Instituto Superior Técnico (IST), Universidade Técnica de Lisboa (UTL)
jose....@ist.utl.pt


_______________________________________________
Haskell-Cafe mailing list
Haskel...@haskell.org
http://www.haskell.org/mailman/listinfo/haskell-cafe

[Kristopher Micinski]

Jose,

So I'm interested to hear you opinion on this as well...

I use Pandoc with Markdown through Hakyll, which allows you to do a
fair amount of cute things that are just really helpful for
maintaining a blog (for example..). But I didn't get this from
reading your github readme: what makes your markup language special?
Could you give an example of how the language is more expressive than
(say) markdown processed through Pandoc (I only mention because it
lets you process LaTeX, very helpful, right...) or something
comparable?

kris

[José Lopes]

Hello Kris,

Thank you for your email.

At this moment, Fmark is not as powerful as Markdown, also because Fmark
just started.
Markdown offers things such as Blockquotes, Lists, Code blocks, links,
emphasis, images, etc.
Fmark does not offer as many features: for now, there are only
paragraphs, headings,
subsections (endless nesting) and footnotes. In the near future, I want
to bolds/italics,
ordered and unordered lists, links, and later on as many elements as
possible :)

The problem with Fmark is also its greatest feature. While other markup
languages
introduce special syntactic characters to give meaning to the document's
elements,
I would like to take a different approach: I want to use characters that
people already
use in document writing to achieve the same result. For example, in
Mediawiki a
heading is some text surrounded by equal signs. But in Fmark a heading
is simply some
text that does not end in a punctuation character, such as period or an
exclamation mark.
I argue that this is a more "natural" approach.

I want to find a natural way of not burdening the user with the task of
having to learn
some special syntax in order to write a document. Instead I want to find
"natural" ways
of writing and use those ways to reconstruct the elements in a document.
Of course,
what is natural is subjective and that is why I want to find a good
tradeoff between
expressiveness and simplicity in the syntax. For example, in Fmark a
footnote is some
text surrounded by square brackets. Maybe you find this natural, maybe
you don't. If a
handful of people defend a more natural way of writing footnotes I want
to implement
the way they say. If there is a more natural way of doing this I want to
find it. But for now
I think square brackets are better than the equal signs or any other
strange syntactic
character such as exclamation marks and so on...

Another thing about Fmark is styles. I want to use fmark personally to
write papers, using
Latex as backend. While experimenting with previous versions of Fmark I
realized that I
could not specify the title, the author, the date, and the abstract.
which are essential in a
paper. I came up with an idea which I think is quite interesting. I
wrote another document
also using Fmark which only had the words "Title", "Author", "Date", and
"Abstract". And
then I combined these two documents together, such that, Fmark
associated title, author,
date and abstract, with the corresponding content. I thought the idea
was interesting
because the content and style documents have both the same structure and
are both
written in Fmark. Of course, there is still a long way to go, in order
to be able to fully
customize a document.

But styles are a good and simple approach, similar to document classes
in Latex: the idea is
to write one document (content) and then use multiple (predefined, user
defined) styles, such
as, article, report, etc, to stylize your document. Another interesting
thing I have been thinking
about (but not implemented yet) is recursion in document styling. In a
way, weaving a style
with content can be compared to matching a regular expression.

Anyway, these are just some key ideas. I see Fmark as a work in progress
and in a way as a
research project, trying to find a natural way of writing documents
while escaping as much
as possible from the syntax of a programming language. I also have a
metagoal with this
project: if my father (the non programming guy) could use it to write
his PhD dissertation,
I would be quite happy :)

If you have any more questions I would be happy to answer.
But if you're interested in using markup languages for blogs perhaps a
HTML backend
in Fmark would be more interesting for you. Although, XML + JavaScript +
CSS is also possible.

Best regards,
José

Instituto Superior Técnico
Technical University of Lisbon

[Richard O'Keefe]

On 18/09/2012, at 3:09 PM, José Lopes wrote:
> Fmark (Friendly Markup) is a very simple markup language without
> syntax and simple but sophisticated document styling, capable of
> producing PDF and XML files.

Do you _really_ mean "without syntax"?

Nope, thought not:

Fmark relies merely on spacing, indentation,
and common punctuation (e.g., periods, exclamation
and interrogation marks) to reconstruct the structure
of your document.

That's SYNTAX.

A markup language without syntax is like a meal without food.

[Ivan Lazar Miljenovic]

Is it possible to override this? What happens if I want a heading of
"This is the greatest Heading Ever!!!!!" ?

"Natural" conventions seem to be to be rather hacky and with lots of
corner cases; I think it's better to define a specific syntax for
markup (e.g. what is the "natural" way of emphasising text?) and stick
to it (though I agree that Markdown has some odd choices; in
particular, the ability to use both _ and * for italics whilst
requiring ** for bold).

--
Ivan Lazar Miljenovic
Ivan.Mi...@gmail.com
http://IvanMiljenovic.wordpress.com

[Richard O'Keefe]

On 18/09/2012, at 3:57 PM, José Lopes wrote:
> The problem with Fmark is also its greatest feature. While other markup languages
> introduce special syntactic characters to give meaning to the document's elements,
> I would like to take a different approach: I want to use characters that people already
> use in document writing to achieve the same result. For example, in Mediawiki a
> heading is some text surrounded by equal signs. But in Fmark a heading is simply some
> text that does not end in a punctuation character, such as period or an exclamation mark.
> I argue that this is a more "natural" approach.

The problem with that is that some people DO end some headings with
a full stop; for them your special syntax is not natural.

>
> I want to find a natural way of not burdening the user with the task of having to learn
> some special syntax in order to write a document.

You haven't found it. What you *have* is very special syntax expressed using
several methods, AND IT IS NOT DOCUMENTED. I have read the examples, and I can
find nothing explaining what the syntax is.

For example, I find indenting subsections rather unnatural and error-prone.
(For example, moving a paragraph from a deep location to a shallow one would
create a new subsection unintentionally.)
Is the amount of indentation fixed? How many levels of subsections are
supported? What if I want to use indentation to express quotation instead?
How do I embed source code? How can you get an example of Fmark in an
Fmark document without having it acted on? I could go on and on with
questions about syntax.

[José Lopes]

Hello Ivan,

I agree with your point: if you want a heading that ends with a punctuation
sign then you cannot do it in Fmark (for now). That gives me something to
think about. However, I will still look for a way that avoids (as much as
possible) special syntax. Do you have any suggestion?

I also agree with you on the "natural" conventions. I want to find a good
tradeoff between syntax and expressiveness. In other words, I want to
avoid as much those "odd choices" you mentioned. For example, I have
been thinking seriously about emphasis and what would be a good way
to do it. So far I could only come up with quotes (either " or '). What do
you think?

Cheers,
José

[José Lopes]

Hello Richard,

When you say "(for) some people (...) you special syntax is not natural"
that's a good thing. I want these people involved in the project. I want
to understand what they find natural in order to weigh the options and
make a proper decision.

On the README file in the github page you will find a brief explanation
of the markup elements. I need to elaborate you that though because
I feel I explain it too fast.
https://github.com/jabolopes/fmark

Let me answer your questions about sections. The amount of indentation
is not fixed. You can use whatever you want. There is also no nesting limit
in Fmark, however, there is a nesting limit in the Latex backend. For now,
quotations, block quotes, source code embedding, etc, are not implemented.
Those will be added in the future.

About embedding a Fmark document in another document. That seems to
be a very cool feature. I will definitely think about it! Maybe you can come
up with a "natural" way of doing it?

Cheers,
José

[Ivan Lazar Miljenovic]

On 18 September 2012 21:53, José Lopes <jose....@ist.utl.pt> wrote:
> Hello Ivan,
>
> I agree with your point: if you want a heading that ends with a punctuation
> sign then you cannot do it in Fmark (for now). That gives me something to
> think about. However, I will still look for a way that avoids (as much as
> possible) special syntax. Do you have any suggestion?
>
> I also agree with you on the "natural" conventions. I want to find a good
> tradeoff between syntax and expressiveness. In other words, I want to
> avoid as much those "odd choices" you mentioned. For example, I have
> been thinking seriously about emphasis and what would be a good way
> to do it. So far I could only come up with quotes (either " or '). What do
> you think?

I think that _emphasis_ is pretty "natural", as is *bold* and possibly
-strikethrough-.

But you _are_ adding in some aspects of markup now.

Using quotes is bad because what happens if you're actually quoting someone? ;-)

--

Ivan Lazar Miljenovic
Ivan.Mi...@gmail.com
http://IvanMiljenovic.wordpress.com

_______________________________________________

[José Lopes]

Hi,

Like I said it's a tradeoff. I will try to use this philosophy as much as
possible. But it's also important not to be fundamentalist. We'll see
as it goes.

I like that strikethrough. You are right about the quotes, but we can
leave " for quoting and use the ' for something else.

Cheers,
José

[Stefan Monnier]

> The problem with that is that some people DO end some headings with
> a full stop; for them your special syntax is not natural.

Markdown/ReST is already using the "no syntax" idea (e.g. compared to
pre-wiki markup such a LaTeX or Texinfo), so he's simply trying to push
this idea further.

I suspect it'll be difficult.

> What if I want to use indentation to express quotation instead?

I think this one is solvable: a paragraph that's more indented than the
previous heading can be considered a quote.


Stefan

[José Lopes]

Hello Stefan,

Thank you for the input.

Cheers,
José

--
José António Branquinho de Oliveira Lopes
Instituto Superior Técnico
Technical University of Lisbon

[Brandon Allbery]

On Tue, Sep 18, 2012 at 1:05 AM, Ivan Lazar Miljenovic <ivan.mi...@gmail.com> wrote:

to it (though I agree that Markdown has some odd choices; in

particular, the ability to use both _ and * for italics whilst
requiring ** for bold).

The odd thing is, I've found that I use those constructs "naturally" (i.e. when simply writing text in a text-only medium) in pretty much exactly the way Markdown uses them.  _italics_ and *italics* are semantically different (I tend to use the former for titles and such) but are both usually italics in my normal usage.  (There's also /italics/ but that is not very widespread.)  I do tend to use # instead of = for headings but that's kinda stolen (and mutated) from a different variety of non-natural markup.

I suspect that when the dust has settled, it'll turn out that Markdown is pretty much right.  Maybe with some tweaking, but largely it represents (and IIRC was derived from) actual in-the-wild textual usage.

--
brandon s allbery                                      allb...@gmail.com
wandering unix systems administrator (available)     (412) 475-9364 vm/sms

[José Lopes]

Hey,

Why do you say that _italics_ and *italics* are semantically different?
What do you mean?

Cheers,
José

[Brandon Allbery]

On Tue, Sep 18, 2012 at 10:34 AM, José Lopes <jose....@ist.utl.pt> wrote:

Why do you say that _italics_ and *italics* are semantically different?
What do you mean?

The parenthetical hinted at it:

(I tend to use the former for titles and such)

There's a convention at work here, namely that italicized book titles and similar are often underlined in plain text — and in plain text without backspaces this is represented as leading and trailing _.  So there's a semantic (i.e. content) difference even though the generated markup is the same.

[Tillmann Rendel]

Hi,

José Lopes wrote in an earlier email:

> I want to find a natural way of not burdening the user with the task of having to learn
> some special syntax in order to write a document.

And then:
> [...] we can leave " for quoting and use the ' for something else.

That sounds like 'some special syntax' to me. Or should it be "some
special syntax"? I can't remember which was for quoting and which was
for something else. Maybe I need to look that up in the Fmark
documentation, again.

Tillmann

[José Lopes]

Hey Tillmann,

That is a good point. What would you suggest for emphasis ?

Cheers,
José

On 18-09-2012 21:00, Tillmann Rendel wrote:
> Hi,
>
> José Lopes wrote in an earlier email:
>> I want to find a natural way of not burdening the user with the task
>> of having to learn
>> some special syntax in order to write a document.
>
> And then:
>> [...] we can leave " for quoting and use the ' for something else.
>
> That sounds like 'some special syntax' to me. Or should it be "some
> special syntax"? I can't remember which was for quoting and which was
> for something else. Maybe I need to look that up in the Fmark
> documentation, again.
>
> Tillmann

--
José António Branquinho de Oliveira Lopes
Instituto Superior Técnico
Technical University of Lisbon

[Richard O'Keefe]

On 19/09/2012, at 12:04 AM, José Lopes wrote:

> Hello Richard,
>
> When you say "(for) some people (...) you special syntax is not natural"
> that's a good thing. I want these people involved in the project. I want
> to understand what they find natural in order to weigh the options and
> make a proper decision.

One important question is how many *scripts* you want to support.
Do you, for example, want to support Greek? There is a Unicode
character for Greek question mark U+037E, some documentation I've
seen says that the preferred character is U+003F. So does ";"
terminate a sentence or not?
How many *languages* do you want to support?
Are you going to support Armenian, where the question mark that
ends a sentence goes not _after the last letter_ of the last
word but _over the last vowel_?

(As someone who only writes in English, I have no trouble with the
answer "only the Latin script, only the English language".)

I don't actually believe that there _is_ a natural convention for
italics, boldface, superscripts, etc. Even _this_ is an artificial
convention going back to mechanical typewriters that could underline
but not change font.

>
> On the README file in the github page you will find a brief explanation
> of the markup elements. I need to elaborate you that though because
> I feel I explain it too fast.
> https://github.com/jabolopes/fmark

I did read the README.md. I disagree that there exists any problem
of manual removal of special characters. If I want to remove XML markup,
I just do unxml foobar.xml >foobar.txt. If I want to convert from one
kind of markup to another, there is pandoc, which is admittedly a touch
buggy, ...

The problem is that the README.md does not in fact explain what any
of the 'markup elements' are.

Let's take one little example, a letter of the kind I used to write
by hand.

123 Copper Road,
Gemstone,
Dunedin,
New Zealand 9999.
31 February 2016.

Dear Schnorer,
I am writing to you because I am tired of frobulating.

Yours without wax,
Dr Strabismus of Utrecht
(whom God preserve!)

How do I tell Fmark about this structure?
How do I get this layout? (The '1' and 'Y' should be in the centre.)

> Let me answer your questions about sections. The amount of indentation
> is not fixed. You can use whatever you want. There is also no nesting limit
> in Fmark, however, there is a nesting limit in the Latex backend. For now,
> quotations, block quotes, source code embedding, etc, are not implemented.
> Those will be added in the future.

These answers belong in the README.me.

>
> About embedding a Fmark document in another document. That seems to
> be a very cool feature. I will definitely think about it! Maybe you can come
> up with a "natural" way of doing it?

The last time I thought I had a natural way to do anything like that
I found that the SGML design was worse broken than I would have believed possible.

[Richard O'Keefe]

On 19/09/2012, at 1:43 AM, Stefan Monnier wrote:

>> The problem with that is that some people DO end some headings with
>> a full stop; for them your special syntax is not natural.
>
> Markdown/ReST is already using the "no syntax" idea (e.g. compared to
> pre-wiki markup such a LaTeX or Texinfo), so he's simply trying to push
> this idea further.

Markdown is very heavy on syntax,
what it is *light* on is specification of what the
syntax actually is. As a result,
I'm aware of three different dialects,
and someone told me about having to reverse
engineer the syntax from a Perl implementation.
As a further result, I cannot write a program to
reliably *generate* Markdown.

>
> I suspect it'll be difficult.

Oh, more power to him for trying.
I just don't think it can be pushed very far.

Oh, there is a really *filthy* hack that could be pulled
for italics, bold face, and so on. Contrary to its original
principles, Unicode includes several copies of ASCII
(see http://unicode.org/charts/PDF/U1D400.pdf):
Mathematical bold,
Mathematical italic,
Mathematical bold italic,
Mathematical script,
Mathematical bold script,
Mathematical fraktur,
Mathematical double struck (blackboard-bold),
Mathematical bold fraktur,
Mathematical sans-serif,
Mathematical sans-serif bold,
Mathematical sans-serif italic,
Mathematical sans-serif bold italic,
Mathematical monospace,
and some similar sets of Greek.
So as long as you don't want strike-through or underlying,
and as long as you don't want italic Cyrillic &c, ...
Too bad if you want a bold italic capital Thorn...

>
>> What if I want to use indentation to express quotation instead?
>
> I think this one is solvable: a paragraph that's more indented than the
> previous heading can be considered a quote.

Ah, but the quotation might not end with a sentence terminator,
so that would be considered a new heading.

[José Lopes]

Hey Richard,

Regarding the languages, I think it is better to start with English
and see how successful Fmark is. There's no point in trying to tackle
lots of languages if we cannot solve the problem for one of them.
But then again, keeping everything Unicode.

In fact, in this matter, I think with Haskell we have an advantage
because things like 'isSpace' and 'isPuncuation' work with Unicode.

I now understand that the README.md is rather incomplete. I will
add some more information about the markup. Although, things
like bold, italics and so on are not implemented yet.

I will also take a look at your letter example. I think the real
challenge here is center alignment. I will think about a
way to solve it in Fmark so we can discuss it later on.

Thank you for the feedback.

Cheers,
José

--
José António Branquinho de Oliveira Lopes
Instituto Superior Técnico
Technical University of Lisbon

[José Lopes]

Hello Richard,

I made a first draft of your letter example.
It is not exactly as you had written but I think you might like it.
It is the combination of a content document (letter) and a style
document (letter.style). You can see the results in the PDF, which
was created using the Latex backend and the letter document
class.

Cheers,
José

[Mario Blažević]

On 12-09-18 07:37 PM, Richard O'Keefe wrote:
>
> On 19/09/2012, at 1:43 AM, Stefan Monnier wrote:
>
>>> The problem with that is that some people DO end some headings with
>>> a full stop; for them your special syntax is not natural.
>>
>> Markdown/ReST is already using the "no syntax" idea (e.g. compared to
>> pre-wiki markup such a LaTeX or Texinfo), so he's simply trying to push
>> this idea further.
>
> Markdown is very heavy on syntax,
> what it is *light* on is specification of what the
> syntax actually is. As a result,
> I'm aware of three different dialects,
> and someone told me about having to reverse
> engineer the syntax from a Perl implementation.
> As a further result, I cannot write a program to
> reliably *generate* Markdown.

Very true. Sadly, this is the case with almost all other Wiki-like
markup schemes out there. They are all implementation-specified. The
only exception I'm aware of is Creole, for which an EBNF grammar exists,
even if it's pretty nasty-looking. A look at that specification makes
one appreciate how badly specified "natural" syntax is.

In my opinion, there is no single natural syntax that can be imposed on
ASCII strings and serve majority of uses. There are many different
syntaxes that feel natural for different uses and different users, and
the best we can hope to achieve would be a way to provide a formal and
readable specification for each of those syntaxes. I've been playing
with one approach in this direction with the concrete-relaxng-parser
package, but it's still early days.

>>
>> I suspect it'll be difficult.
>
> Oh, more power to him for trying.
> I just don't think it can be pushed very far.
>
> Oh, there is a really *filthy* hack that could be pulled
> for italics, bold face, and so on. Contrary to its original
> principles, Unicode includes several copies of ASCII
> (see http://unicode.org/charts/PDF/U1D400.pdf):
> Mathematical bold,
> Mathematical italic,
> Mathematical bold italic,
> Mathematical script,
> Mathematical bold script,
> Mathematical fraktur,
> Mathematical double struck (blackboard-bold),
> Mathematical bold fraktur,
> Mathematical sans-serif,
> Mathematical sans-serif bold,
> Mathematical sans-serif italic,
> Mathematical sans-serif bold italic,
> Mathematical monospace,
> and some similar sets of Greek.

Thank you for sharing this hack. It's very amusing.




--
Mario Blazevic
mbla...@stilo.com
Stilo International

This message, including any attachments, is for the sole use of the
intended recipient(s) and may contain confidential and privileged
information. Any unauthorized review, use, disclosure, copying, or
distribution is strictly prohibited. If you are not the intended
recipient(s) please contact the sender by reply email and destroy
all copies of the original message and any attachments.