Legal Team/Templatizing/tags-matching

From SPDX Wiki
< Legal Team‎ | Templatizing
Revision as of 00:36, 31 March 2016 by Goneall (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

SPDX License List Matching Guidelines and tag name

XML templatizing tag values and nesting rules

The general structure of the MASTER xml files (located in /src) is as follows:

<SPDX name="full name" identifier="SPDX short identifier" osi-approved="true/false">
  <urls>
    <url>URL for license</url>
    <url>additional URL for license</url>
  </urls>
  <notes>notes on license, if any</notes>
  <header>standard license header, if any</header>
  <license>
    <title>license name or title (Guideline 10) </title>
    <copyright>Copyright notice (Guideline 9) </copyright>
    <body> Substantive text of license - most matching guidelines apply here
      <list>
        <li><b>a bullet (Guideline 7) </b>text of the list item</li>
      </list>
      <p>some more text</p> 
    </body>   
    <optional>some text that is omitable (Guideline 2) </optional>
  </license>
</SPDX>

The tags are defined as follows:

  • SPDX tag: root element tag; has attributed to specify full license name, short identifier, and OSI-approved (if the license is OSI approved, this field/tag will indicate "yes", otherwise, left blank) [Equivalent property names in the published format: licenseId, isOsiApproved, name]
  • URLs/URL tag: each url for the official text of the license or exception [Note: This is the seeAlso tag in the published format]
  • Notes tag: factual info about the license, e.g. date of release, license has been deprecated, links to official translations [Note: this is the comment tag in the published format]
  • Header tag: the short form of a license that references the full license text. This is a separate entity, not a part of the license proper [Note: this is the standardLicenseHeader tag in the published fromat]
  • License tag: the root element that contains the license data [Same tag name as the published format]
  • Title tag: denotes an optional (as per Matching Guideline 10) section of the license content, with the implication that it contains the title/heading content that prefaces the substantive text [There is no equivalent tag in the published format - suggest title, lower case to be consistent with Dublin Core spec]
  • Copyright tag: denotes an optional (as per Matching Guideline 9) section of the license content, with the implication that it contains copyright notices *that apply to the code under license*; this is explicitly *not* intended to include copyright notices that apply to the license itself
  • Body tag: substantive text of the license itself [There is no equivalent tag in the published format, there is copyrightText property name used in other parts of the spec]
  • B tag: denotes a bullet or section marker whose style and formatting is unimportant to the substantive content of the license text as per Matching Guideline 7 [There is no equivalent tag in the published fomart - B is fine]
  • Optional tag: denotes otherwise optional (as per Matching Guildeline 2) text content, such as appendices explaining how to use a license and/or defining the license's header form [This is OK - could translate to beginOptional/endOptional]
  • alt tag: denotes a small portion of the text that is specific to the content being licensed, such as the name of the copyright holder. The "name" attribute can be ignored for matching purposes, but if present should specify a unique identifier that the actual matched content (when a template is applied to some target text for matching) should be stored in. The "match" attribute specifies a regular expression that governs what text is suitable/allowed to match in lieu of the original text content [Currently, this is "var", but alt should be OK]

The general structure of the MATCHING FORMAT (output/build product) xml files (not yet produced) is as follows:

<license name="Full license name" identifier="Identifier">
  <header>Header version of license</header>
  <body>
    <title>Optional title text</title>
    <copyright>Copyright notice</copyright>
    Substantive license text
    <b>bullet</b> more text...
    Sometimes the text will include <alt name="foo" match="bar">original text</alt>
    The only difference between the master xml file and the matching format xml files is the <syn identifier="synonym">synonym</syn> tag.
    <optional>Other optional text such as appendices</optional>
  </body>
</license>

This format is almost identical to the master files, with the exception that synonyms are automatically identified and marked as part of a build process. This allows a separate list of valid synonyms to be maintained and referenced without needing to find and mark all synonyms by hand or change every xml file if the synonym list in altered.

  • syn tag: denotes meaningful substantive text that may have variations in spelling (regional or otherwise), for which alternate spellings are valid and have identical meaning. A separate file should be curated that lists all applicable synonyms. The "identifier" attribute specifies which set of synonymous words may match, and the tag wraps the original text as given in the license proper.

Specific use cases

Display the original text content of the license: extract the inner text content of the <body> tag, stripping all XML.

Display the substantive license text of the license: extract the inner text content of the <body> tag, removing all optional tag sections.

Generate a regular expression to be used for identifying a license: within the body tag, replace all instances of the <b< tag with a suitable regular expression to match bullets. Replace all instances of the <alt> tag with the regular expression specified by the "match" attribute. Replace all instances of the <syn> tag with a regular expression that matches only the allowed set of synonyms. Flatten whitespace and match case-insensitively. Target text will also need to be normalized by flattening whitespace. Alternately, make the regular expression all lowercase, and normalize all target text to lowercase.

License header text should be matched separately/independently of the license body.

SPDX License List Matching Guidelines (draft revisions)

SPDX License List Matching Guidelines, v2.0

This document provides guidelines to be used for the purposes of matching licenses and license exceptions against those included on the SPDX License List. There is no intent here to make a judgment or interpretation, but merely to ensure that when one SPDX user identifies a license as "BSD 3-clause," for example, it is indeed the same license as what someone else identifies as "BSD 3-clause" and the same license as what is listed on the SPDX License List. Examples of how to apply some of the matching guidelines to a license or exception are provided via templates.  Templates are comprised of technical markup within the master license text file to provide further or specific guidance to SPDX document creators or tool makers.  Not all licenses or exceptions will have templates.

1. How These Guidelines Are Applied

1.1 Purpose:  To ensure consistent results by different SPDX document creators when matching license information that will be included in the License Information in File field. SPDX document creators or tools may match on the license or exception text itself, the official license header, or the SPDX License List short identifier.

1.1.1 Guideline: Official License Headers  The same matching guidelines used for license and exception text apply to the official license headers. Where applicable, an official license header template file with markup is included with the SPDX License List. Official license headers are defined by the SPDX License List as specific text specified within the license itself to be put in the header of files. (see <a href="http://spdx.org/spdx-license-list/license-list-overview">http://spdx.org/spdx-license-list/license-list-overview</a> for more info).

2. Substantive Text

2.1 Purpose:  To ensure that when matching licenses and exceptions to the SPDX License List, there is an appropriate balance between matching against the substantive text and disregarding parts of the text that do not alter the substantive text. Further guidelines of what can be disregarded or considered replaceable for purposes of matching are listed below here and in the subsequent specific guidelines. A conservative approach is taken in regards to rules about disregarded or replaceable text.

2.1.1 Guideline:  Verbatim Text  License and exception text should be the same verbatim text (except for the guidelines stated here). The text should be in the same order, e.g., differently ordered paragraphs would not be considered a match.

2.1.2 Guideline:   No Additional Text Matched text should only include that found in the vetted license or exception text. Where a license or exception found includes additional text or clauses, this should not be considered a match.

2.1.3 Guideline:  Replaceable Text  Some licenses include text that refers to the specific copyright holder or author, yet the rest of the license is exactly the same as a generic version. The intent here is to avoid the inclusion of a specific name in one part of the license resulting in a non-match where the license is otherwise an exact match (e.g., the third clause and disclaimer in the BSD licenses, or the third, fourth, and fifth clauses of Apache 1.1). In these cases, there should be a positive license match.

Text that can be considered replaceable for matching purposes is indicated in the SPDX License List template with mark-up and in the corresponding HTML pages with colored text. The text indicated as such can be replaced with similar values (e.g., a different name or generic term; different date) and still be considered a positive match. This rule also applies to text-matching in official license headers (see Guideline #1).

2.1.4 Guideline:  Omitable Text  Some licenses have text that can simply be ignored. The intent here is to avoid the inclusion of certain text that is superfluous or irrelevant in regards to the substantive license text resulting in a non-match where the license is otherwise an exact match (e.g., directions on how to apply the license or other similar non-substantive exhibits). In these cases, there should be a positive license match.

Text that can be considered omitable for matching purposes is indicated in the SPDX License List template with mark-up and in the corresponding HTML pages with colored text. The license should be considered a match if the text indicated is present and matches OR the text indicated is missing altogether.

3. Whitespace

3.1 Purpose:   To avoid the possibility of a non-match due to different spacing of words, line breaks, or paragraphs.

3.1.1 Guideline:  All whitespace should be treated as a single blank space. Templates do not include markup for this guideline.

4. Capitalization

4.1 Purpose:  To avoid the possibility of a non-match due to lower case or upper case letters in otherwise the same words.

4.1.1 Guideline:  All upper case and lower case letters should be treated as lower case lettersTemplates do not include markup for this guideline.

5. Punctuation

5.1 Purpose:  Because punctuation can change the meaning of a sentence, punctuation needs to be included in the matching process. License template files do not include markup for this guideline.

5.1.1 Guideline:  Punctuation  Punctuation should be matched, unless otherwise stated in these guidelines.

5.1.2 Guideline:  Hyphens, Dashes  Any hyphen, dash, en dash, em dash, or other variation should be considered equivalent.

5.1.3 Guideline:  Quotes  Any variation of quotations (single, double, curly, etc.) should be considered equivalent.

6. Code Comment Indicators

6.1 Purpose:  To avoid the possibility of a non-match due to the existence or absence of code comment indicators placed within the license text, e.g. at the start of each line of text.

6.1.1 Guideline:  Any kind of code comment indicator or prefix which occurs at the beginning of each line in a matchable section should be ignored for matching purposes. Templates do not include markup for this guideline.

7. Bullets and Numbering

7.1 Purpose:  To avoid the possibility of a non-match due to the otherwise same license using bullets instead of numbers, number instead of letter, or no bullets instead of bullet, etc., for a list of clauses.

7.1.1 Guideline:  Where the template specifies a bullet (with the <b> tag), the target text may include bullets or bullet-like text [TODO: specify precisely what counts, possibly provide a regular expression]. Templates include markup for this guideline.

8. Varietal Word Spelling

8.1 Purpose:  English uses different spelling for some words. By identifying the spelling variations for words found or likely to be found in licenses, we avoid the possibility of a non-match due to the same word being spelled differently. This list is not meant to be an exhaustive list of all spelling variations, but meant to capture the words most likely to be found in open source software licenses.

8.1.1 Guideline:  The SPDX templates are published alongside an XML file containing a list of valid synonyms. Where a template specifies the <syn> tag, any word in this list with the same identifier may be matched. Templates include markup for this guideline.

9. Copyright Symbol

9.1 Purpose:  By having a rule regarding the use of "©", "(c)", or "copyright", we avoid the possibility of a mismatch based on these variations.

9.1.1 Guideline:  Different forms of the copyright identifier (e.g. "©", "(c)", or "Copyright") should be considered equivalent and interchangeable. They are included in the synonym list and marked up in the template files accordingly. Templates include markup for this guideline.

10. Copyright Notice

10.1 Purpose:  To avoid a license mismatch merely because the copyright notice (usually found above the actual license or exception text) is different. The copyright notice is important information to be recorded elsewhere in the SPDX file, but for the purposes of matching a license to the SPDX License List, it should be ignored because it is not part of the substantive license text.

10.1.1 Guideline:  You may ignore the contents of the <copyright> tag for the purposes of matching. Templates include markup for this guideline.

11. License Name or Title

11.1 Purpose:  To avoid a license mismatch merely because the name or title of the license is different than how the license is usually referred to or different than the SPDX full name. This also avoids a mismatch if the title or name of the license is simply not included.

11.1.1 Guideline:  You may ignore the contents of the <title> tag for the purposes of matching. Templates include markup for this guideline.

12. Extraneous Text At the End of a License

12.1 Purpose:  To avoid a license mismatch merely because extraneous text that appears at the end of the terms of a license is different or missing. This also avoids a mismatch if the extraneous text merely serves as a license notice example and includes a specific copyright holder's name.

12.1.1 Guideline:  You may ignore the contents of the <optional> tag for the purposes of matching. Templates include markup for this guideline.

Notes

There remains one slightly computer-hard problem to solve with regards to matching a template to a license: code comment indicators. This cannot be solved by the SPDX templates themselves, since the text that needs to be modified is the target text being matched against, not the markup that specifies the matching process.

I've changed my mind on the subject of formatting the templates in lowercase and stripping out whitespace: these can be done readily and easily by a program, and preemptively applying them in the build process destroys information unnecessarily. I've also decided to keep the master and matching formats equivalent in all other ways.