TIP 3: TIP Format

Login
Author:         Andreas Kupries <[email protected]>
Author:         Donal K. Fellows <[email protected]>
State:          Accepted
Type:           Process
Vote:           Done
Created:        14-Sep-2000
Post-History:
Obsoleted-By:   467

Abstract

This TIP is a companion document to the TIP Guidelines [2] and describes the structure and formatting to use when writing a TIP.

Rationale

The major goals of this document are to define a format that is

The latter is important because non-acceptance essentially means that the TIP process will be stillborn. This not only means basically plain text without much markup but also that we should reuse formats with which people are already acquainted.

As the concept of TIPs borrows heavily from Python's PEPs http://python.sourceforge.net/peps/ their definition on how to structure and format a PEP was reviewed for its suitability of use by the TCT and the community at large.

The major points of the format are:

A header section like is used in mail or news is something people are acquainted with and fulfils the other criteria too. In addition it is extendable. Using indentation to convey semantic and syntactic information on the other hand is something Pythonistas are used to but here in the Tcl world we are not to the same extent.

Looking at bit closer to home we find the Tcl'ers Wiki http://wiki.tcl.tk/

It does use a plain text format with some very light formatting conventions to allow things like links, images, enumerated and itemized lists.

Given the rather high acceptance of this site by the community using its format should be beneficiary to the acceptance of TIPs too.

It is therefore proposed to use a combination of a header in mail/news style together with a body employing a slightly extended/modified Wiki format (mostly backward compatible) as the format for TIPs. This proposed format is specified in detail below.

Note that the use of TAB characters within a TIP is discouraged (but permitted) as some mailers (notably Outlook Express) make a mess of them. Please be considerate and avoid their use...

Rejected Alternatives

But before we specify the format a (short) discussion of possible alternatives and why they where rejected.

There were three primary competitors to the format specified below, these are SGML/XML, HTML and a markup based upon plain text with embedded tcl-commands, for example like ... [section Abstract] ...

Header Format

The general format of the header for a TIP is specified in RFC 822 http://www.rfc-editor.org/rfc/rfc822.txt . This leaves us to define and explain the keywords, their meaning and their values. The following keywords are required, and unless otherwise stated, should occur exactly once:

TIP: The number of the TIP as assigned by the TIP editor. Unchangeable later on.

Title: Defines the title of the document using plain text. May change during the discussion and review phases.

Version: Specifies the version of the document. Usually something like $Revision: 1.8 $. (Initially $Revision: 1.8 $ should be used, which is then changed by the version control to contain the actual revision number.

Author: Contact information (email address) for each author. The email address has to contain the real name of the author. If there are multiple authors of the document, this header may occur multiple times (once per author.) The format should be approximately like this: Firstname Lastname

State: Defines the state the TIP is currently in. Allowed values are Draft, Active, Accepted, Deferred, Final, Rejected and Withdrawn. This list will be influenced by the finalization of the workflow in [2].

Type: The type of the TIP. Allowed values are Process, Project and Informative. See [2] for more explanations about the various types.

Vote: The current state of voting for the TIP. Allowed values are Pending, In progress, Done and No voting. The latter is used to indicate a TIP which doesn't require a vote, for example [1].

Created: The date the TIP was created, in the format dd-mmm-yyyy. mmm is the (English) short name of the month. The other information is numerical. Example: 14-Sep-2000

All numeric dates, though more easily internationalised, are not used because the ordering of particularly the month and day is ambiguous and subject to some confusion between different locales. Unix-style timestamps are unreadable to the majority of people (as well as being over-precise,) and I ([email protected]) don't know ISO 8601 well enough to be able to comment on it.

Post-History: A list of the dates the document was posted to the mailing list for discussion.

Tcl-Version: This indicates the version of Tcl that a Project TIP depends upon (where it is required.) Process and Informative TIPs must not have this keyword.

The following headers are optional and should (unless otherwise stated) occur at most once:

Discussions-To: While a TIP is in private discussions (usually during the initial Draft phase), this header will indicate the mailing list or URL where the TIP is being discussed.

Obsoletes: Indicates a TIP number that this TIP renders obsolete. (Thanks to Joel Saunier [email protected] for suggesting this!)

Obsoleted-By: Indicates a TIP number that renders this TIP obsolete. (Thanks to Joel Saunier [email protected] for suggesting this!)

Keywords: A comma-separated list of keywords relating to this TIP, to facilitate automated indexing and improve search engine results.

The following headers are proposed (by Donald G. Porter [email protected]) but not currently supported:

Sponsor: A TCT member that is sponsoring this TIP. May occur multiple times, once per sponsor.

Supporter: A person (not necessarily a TCT member) who is supporting this TIP. May occur multiple times, once per supporter.

Objector: A person (not necessarily a TCT member) who is opposed to this TIP. May occur multiple times, once per objector.

Body Format

The body of a TIP is split by visually blank lines (i.e. lines containing nothing other than conventional whitespace) into units that will be called paragraphs. Each paragraph is in one of the following forms.

If the paragraph consists of exactly four minus symbols "----" then it is a separator paragraph and should be rendered as a horizonal rule.

If the paragraph consists of a vertical bar "|" followed by text, then it is a verbatim paragraph. The bar will be stripped from the front of each line and the rest of the text will be formatted literally. Tab characters will be expanded to 8-character boundaries. (Note that this is completely incompatible with the Tcl'ers Wiki.)

If the paragraph consists of one or more tildes "~" (which may be space-separated) followed by text, then it is a section heading. The text following is the name of the section. In the name of good style, the section heading should have its significant words capitalised. The number of tildes indicates whether this is a section heading, a subsection heading or a subsubsection heading (one, two or three tildes respectively.)

If the paragraph consists of the sequence "#index:" followed by some optional text, then it is a request to insert an index. The text following (after trimming spaces) indicates the kind of index desired. The default is a "medium" index, and fully compliant implementations should support "short" (expected to contain less detail) and "long" (expected to contain all header details plus the abstract) as well. Support for other kinds of indices is optional.

If the paragraph consists of the sequence "#image:" followed by some text, then it is a request to insert an image. The first word of the following text is a reference to the image, and the other words are an optional caption for the image (in plain text.) Image references that consist of just letters, numbers, hyphens and underscores are handled specially by the current implementation, which can map them to the correct media type for its current output format (assuming it has a suitable image in its repository.)

All other paragraphs that start with a non-whitespace character are ordinary paragraphs.

If a paragraph starts with a whitespace character sequence (use three spaces and keep the whole paragraph on a single line if you want compatability with the Tcl'ers Wiki,) a star "*" and another whitespace character, it is an item in a bulleted list.

If a paragraph starts with a whitespace character sequence, a number, a full stop "." and another whitespace character, it is an item in an enumerated list. If the number is 1 then the number of the item is guessed from the current list context, and any other value sets the number explicitly. If you want compatability with the Tcl'ers Wiki, make the initial whitespace sequence be three spaces, the number be 1, and keep the whole paragraph on a single line.

If a paragraph starts with a whitespace character sequence, some text (that includes no tabs or newlines but can include spaces), a colon and another whitespace character, then it is an item in a descriptive (a.k.a. definition) list. The item being described cannot contain advanced formatting (including any kind of emphasis) because this is not supported by all formats that a TIP may be viewed in.

If a paragraph does not start with a whitespace character sequence, a greater than symbol ">", and then another whitespace character, it is also an ordinary paragraph. (Note that this is completely incompatible with the Tcl'ers Wiki.)

Where a paragraph does begin with the sequence described in the preceding paragraph, it is a nested list item (if the paragraph contained is a list item) or a subsequent paragraph (if the paragraph contained is an ordinary paragraph.) If there's no suitable "enclosing" list context (i.e. if the preceding paragraph was not part of a list) the paragraph will be a quotation instead. (The rules for these continuation paras seem complex at first glance, but seem to work out fairly well in practise, especially since they are only rarely used.)

Within the body text of a (non-verbatim) paragraph, there are two styles of emphasis:

The two emphasis styles should not be nested. Special URLs of the form tip:tipnumber are expanded into full URLs to the given TIP through the current formatting engine (where applicable.) References of the form [tipnumber] are also expanded as links to the given TIP, but are not displayed as URLs (the expansion is format dependent, of course.) Doubled up square brackets are converted into matching single square brackets. Email addresses (of the form email@address) and ordinary URLs in single square brackets might also be treated specially.

The first paragraph of the body of any TIP must be an abstract section title ("~Abstract" or "~ Abstract"), and the second must be an ordinary paragraph (and should normally be just plain text, to make processing by tools easier.)

You can compare these rules with those for the Tcl'ers Wiki which are described at http://wiki.tcl.tk/14.html, with the following modifications:

  1. The text for an item in an itemized, enumerated or tagged list can be split over multiple physical lines. The text of the item will reach until the next empty line.

  2. All paragraphs must be split with whitespace. This is a corollary of the above item.

  3. A paragraph starting with the character ~ is interpreted as a section heading. Consequently it should be very short so that it renders onto a single line under most circumstances.

  4. A full verbatim mode is added. Any line starting with the bar character is reproduced essentially verbatim (the bar character is removed). This allows embedding of code or other texts containing formatting usually recognized as special by the formatter without triggering this special processing. This applies especially to brackets and the hyperlinking they provide and their role in tcl code. This is used in preference to the whitespace rule of the Tcl'ers Wiki which is potentially far more sensitive. Our rule makes it extremely obvious what lines are verbatim, and what those lines will be rendered as.

  5. Only one style of emphasis within paragraphs is supported. Having multiple emphasis styles (italic and bold) not only fails to carry across well in all media, but also makes for confusion on the part of authors and is more difficult to write renderers for too.

  6. Images are only supported in a limited way, since under HTML the support for images varies a lot more than most people would like to think, and the concept of an inline image can vary quite a lot between different rendered formats too.

Reference Implementation

A reference renderer was created by Donal Fellows [email protected] and is installed (as a behind-the-scenes rendering engine) on a set of TIP documents http://www.cs.man.ac.uk/fellowsd-bin/TIP with the source code to the rendering engine being available http://sf.net/projects/tiprender/

Note that this code does support nested lists and multi-paragraph items, but this is experimental right now. Examples are presented behind the code itself.


Examples

This document itself is an example of the new format.

Examples for nested lists, multi-paragraph items in list's, and quotations.

Here is the source (itself a demonstration of verbatim text)

 * This is a paragraph

 > * This is an inner paragraph
     that goes onto two lines.

 > > * This one's even further in!

 > > * So's this one.

 > * Out again

 > > And a second paragraph here...

 > * Yet another item.

 * Outermost level once more.

 1. Enumerate?

 > 1. Deeper?

 2. Out again?

 list item: body text that is relatively long so that we can tell
   that it laps round properly as a paragraph even though this takes a
   ridiculous amount of text on my browser...

| VERB IN LIST?

 > nested: body

Top-level paragraph once more.

 > A quotation from someone famous might be rendered something like
   this.  As you can see, it is inset somewhat from the surrounding
   text. - ''Donal K. Fellows <[email protected]>''

And back to the top-level yet again.  Now we show off both ''italic''
and '''bold''' text.

----

and the rendered result

list item: body text that is relatively long so that we can tell that it laps round properly as a paragraph even though this takes a ridiculous amount of text on my browser...

	 VERB IN LIST?

 > nested: body

Top-level paragraph once more.

A quotation from someone famous might be rendered something like this. As you can see, it is inset somewhat from the surrounding text. - Donal K. Fellows [email protected]

And back to the top-level yet again. Now we show off both italic and bold text.


Examples of index generation and image paragraphs.

Here is the code

#index:

#index:short

#index: long

#image:3example This is a test caption

This is an example long TIP reference tip:3 that should be expanded in
a renderer-specific way...

This is an example non-reference - ''index[[3]]'' - that should not
be rendered as a link (to this document or anywhere else) at all.
Note that the dashes in the previous sentence (with whitespace on
each side) are candidates for rendering as long dashes (em-dashes) on
output-media which support this.

 Supported URLs: should be http, https, mailto, news, newsrc, ftp and
   gopher.  Test here...

 > HTTP URL - http://purl.org/thecliff/tcl/wiki/

 > HTTP URL in brackets - [http://wiki.tcl.tk]

 > HTTPS URL - https://sourceforge.net/

 > FTP URL - ftp://src.doc.ic.ac.uk/packages/tcl/tcl/

 > NEWS URL - news:comp.lang.tcl

 > MAILTO URL - mailto:[email protected]?subject=TIP3

 > Others (might not be valid links!) - gopher://info.mcc.ac.uk,
   newsrc:2845823825

and here is the rendered result.

#index:

#index:short

#index: long

This is a test caption

This is an example long TIP reference tip:3 that should be expanded in a renderer-specific way...

This is an example non-reference - index[3] - that should not be rendered as a link (to this document or anywhere else) at all. Note that the dashes in the previous sentence (with whitespace on each side) are candidates for rendering as long dashes (em-dashes) on output-media which support this.

Supported URLs: should be http, https, mailto, news, newsrc, ftp and gopher. Test here...

HTTP URL - http://purl.org/thecliff/tcl/wiki/

HTTP URL in brackets - http://wiki.tcl.tk

HTTPS URL - https://sourceforge.net/

FTP URL - ftp://src.doc.ic.ac.uk/packages/tcl/tcl/

NEWS URL - news:comp.lang.tcl

MAILTO URL - mailto:[email protected]?subject=TIP3

Others (might not be valid links!) - gopher://info.mcc.ac.uk, newsrc:2845823825


Examples of sections and subsections

~Section Heading

Section text

~~Subsection Heading

Subsection text

~~~Subsubsection Heading

Subsubsection text

which renders as...

Section Heading

Section text

Subsection Heading

Subsection text

Subsubsection Heading

Subsubsection text


Copyright

This document has been placed in the public domain.