The convention by which names are assigned to Tipitaka files at Access to Insight was implemented in June 2006. It emerged from more than two years of discussions between John Bullitt, Michael Olds, Alex Genaud, and Hugo Gayosso. Several others contributed helpful ideas on the now-defunct ATI technical blog, including John Gill, E.M., Donnovan Knight, Ben N, and Robert.

Contents

Motivation for the convention

The problem

The naming and organization of Tipitaka files on websites across the Internet tends to be haphazard. To illustrate, consider the following suttas from the Majjhima Nikaya that turned up in a recent »Google search for majjhima sutta:

This simple experiment reveals two striking facts:

1. Filenames are inconsistent.
   Some sites name a sutta file using a Pali name (028-mahahatthipadopama-sutta-e1.htm); some use the local language's translation of the Pali title (sloni-stopa.htm); some use a local index number (95.htm); some use uppercase nikaya abbreviations (e.g., MN); some use lowercase; some use hyphens and underscores. And so on.

2. Directory hierarchies are inconsistent.
   Moreover, they are rarely laid out in a way that reflects the structure of the Tipitaka itself. For example: some sites pour all their suttas into one big directory called "tipitaka" (/tipitaka/mn28.htm); others place them in a "sutta" directory (/sutta/MN28.htm); others refine this a little further, by placing each nikaya inside a "canon" directory (/budizam/canon/majjhima/mn28.html); while others place the nikayas under a "sutta pitaka" directory (/canon/sutta/majjhima/mn028-tb0.html).

Why we should care

It is quite natural for a busy webmaster to improvise a filing system that fulfills a site's immediate needs, one that quickly gets the job done of making the site's files accessible to its users. This is the strategy that Buddhist sites (Access to Insight included) have usually followed. But, as any experienced webmaster knows, as the number of files in a site's collection grows, the problem of intelligently managing all the hyperlinks between files can escalate at an alarming rate. 1 Any webmaster hoping to host a substantial collection of Tipitaka files must be prepared to provide hyperlinks between many thousands of files. Adding new files to a growing collection is only practical if the site adheres to a consistent and logical file-naming system.

There are other compelling reasons for consistent Tipitaka filenames. As Buddhist students, practitioners, and scholars study the online Tipitaka texts, they naturally exchange Tipitaka files with each other via e-mail or via postings on their websites. If you were to send me a file named 95.htm along with an e-mail that said, "Here's that sutta we talked about two weeks ago," what am I to make of it? I would have to open it and read it before I could know where to file it. If the file instead had a more meaningful name like mn-095-thanissaro.html then I could tell at a glance that this was Thanissaro's translation of MN 95 and I would know instantly where to drop it in my Tipitaka collection. Moreover, if I am trying to locate MN 95 on a website that is poorly indexed and whose filenames are poorly chosen, there is no way to know a priori where to find that sutta. This makes it extremely difficult to install useful hyperlinks between sites, as the webmaster must first decode the target site's opaque filing system.

Large sites and sites that serve as distribution source points for other sites should be concerned about these issues if they hope to continue to provide well organized and richly cross-referenced collections of Tipitaka texts.

General principles

Restrict the character set

Filenames shall be constructed from a subset of the » World Wide Web Consortium's "unreserved characters" for URIs. In particular, our character set is this:

a b c d e f g h i j k l m n o p q r s t u v w x y z 0 1 2 3 4 5 6 7 8 9 - (hyphen) . (dot)

Please note that we exclude:

• the tilde (~) [because it has a special meaning in Unix C-style shell scripts;
• any uppercase characters [to make URIs easier to type and to remember]
• spaces or tabs [because (1) URIs would have to encode them as %20; (2) filenames with whitespace are hard to type at the command line.]
• Underscores (_) [because they look like spaces in hyperlinked file lists. Example: spot the underscores and the spaces in this file name: an01_03 -123_ xyz.html.]

Keep filenames short

Filenames shall be restricted to no more than 25 characters, plus the extension. For HTML files, either the .html or .htm extension may be used (ATI prefers .html). The maximum filename length is thus 30 characters.

One sutta, one translator, one file

Filenames shall describe a specific sutta. Because of the brevity of suttas in some sections of the canon, however, suttas may sometimes be bundled together into a single file (e.g., AN I, Thag 1, and Thig 1). In some cases, the filename may describe a range of suttas or excerpts. The texts in a file shall always be from a single translator. If a translator has more than one translation of a sutta, all such translations shall be included in the same file.

Filenames reflect the text's location in the canon

Each filename shall be carefully chosen so that, when a file is separated from its enclosing directory (as, for example, when sent as an e-mail attachment), anyone can tell at a glance from where in the canon it came.2

Filenames accurately reflect the contents of the file

Filenames shall include standardized information about (1) which texts the file contains, and (2) the identity of the translator. Filenames may someday be extended to include other information — for example, about the language of the translations, if other than English.

The structure of a filename

To make filenames machine-readable, they are structured to consist of three, four, or five dot-separated data fields, terminated by an extension (EXT). The possible fields are:

VOLUME

The VOLUME field is present in all files. It consists of a two-, three-, or four-character lowercase alphanumeric field. The following VOLUMEs are allowed:

	dn
	mn
	sn01 sn02 sn03 ... sn55 sn56 (one volume per samyutta)
	an01 an02 an03 ... an10 an11 (one volume per nipata)
	khp dhp ud iti snp vv pv thag thig nm
	miln
	mv cv (Vinaya)

Files from a given VOLUME are stored in a directory with the same name as the VOLUME. Thus, files from VOLUME dn are stored in the directory named dn; files from VOLUME an06 are stored in the directory named an06; etc.

The following are not yet implemented on ATI, and will be added on an as-needed basis: pk jat nc ps ap bv cp nett petk (Sutta Pitaka); sv (Vinaya Pitaka); dhs vbh kvu pug dhk yam pat (Abhidhamma Pitaka).

CHAPTER

The CHAPTER field is not present in all files. It corresponds to the vagga or nipata in a given VOLUME. It consists of a one-digit non-padded field or a two-digit zero-padded field.

The following CHAPTERs are recognized:

   VOLUME      CHAPTER
     cv        01 02 03 ... 11 12
     mv        01 02 03 ... 09 10

     ud        1 2 3 4 5 6 7 8
     iti       1 2 3 4
     snp       1 2 3 4 5
     vv        1 2 3 4 5 6 7
     pv        1 2 3 4
     thag      01 02 03 ... 20 21
     thig      01 02 03 ... 15 16
     miln      1 2 3 4 5 6 7

TEXT

The TEXT field is present in all files. In most cases it corresponds to the basic unit of text in the Sutta Pitaka: the sutta. It consists of a one-, two-, or three-digit zero-padded subfield, optionally followed by:

1. the letter x to indicate an excerpt;
2. a hyphen followed by another digit subfield of the same width, to indicate an uninterrupted range of texts;
3. a hyphen, followed by another digit subfield of the same width, followed by the letter x, to indicate a broken series of excerpts.

The possible values in the first subfield of the TEXT field are as follows:

   VOLUME      Possible TEXT values
     cv        01 02 03 ... 
     mv        01 02 03 ... 

     dn         01 02 ... 33 34
     mn         001 002 003 ... 151 152
     sn (all)   001 002 003 ...
     an (all)   001 002 003 ...
     khp        1 2 3 ... 8 9
     dhp        01 02 03 ... 25 26
     ud         01 02 03 ... 
     iti*       001 002 003 ... 110 111
     snp        01 02 03 ... 
     thag       01 02 03 ... 
     thig       01 02 03 ... 
     
        
         *TEXT numbering runs consecutively across the chapters.

The optional part of the TEXT contains more specific information about the contents of the file, as described above. For example:

Filename	What it contains
an10.001x.abcd.html	An excerpt from AN 10.1
an01.001-010.abcd.html	Suttas AN 1.1-10 (complete)
an01.001-010x.abcd.html	Suttas AN 1.1-10 (excerpts)

SECTION

SECTION is a one-digit field that is used to further subdivide certain TEXTs: the Mv and Cv in the Vinaya (following IB Horner's numbering), and the longer suttas in DN.

SOURCE

The SOURCE field is present in all files. It contains a four-byte translator code that stands for the name of the translator(s) of the given text. For example: nyam=Ñanamoli Thera; than=Thanissaro Bhikkhu; irel=John D. Ireland. Librarians and webmasters are encouraged to use translator codes from Access to Insight's pool of Reserved Translator Codes. Eventually, the need may emerge for some organizational structure to oversee and coordinate assignment of these codes — especially when managing translations into other (non-English) languages. But this goes far beyond the scope of what I can do.

Additional SOURCE sub-fields may be needed in the future. All such sub-fields shall be appended after the translator code. If specific languages are to be incorporated in the future, language codes should observe the three-byte ISO 639.2 standard.

Directory hierarchy

tipitaka/
   	
         vin/
              cv/  mv/  sv/ 
         
         dn/
         
         mn/
         
         sn/
              sn01/   sn02/   sn03/   ...   sn55/   sn56/
         
         an/
              an01/   an02/   an03/   ...   an10/   an11/
         
         kn/
              khp/  dhp/  ud/  iti/  snp/  vv/  pv/  thag/  thig/ 
              jat/  nm/   nc/  ps/  ap/   bv/  cp/
              nett/ petk/  miln/
         
         abhi/
              dhs/  vbh/  kvu/  pug/  dhk/  yam/  pat/

Filenames for each Tipitaka section

Mahavagga (Mv)

Note: ATI follows IB Horner's PTS numbering (Book of the Discipline, Part 4).

Form:	VOLUME.CHAPTER.TEXT.SECTION.SOURCE.EXT
VOLUME:	mv
CHAPTER:	01 02 03 ... 11 12
TEXT:	The range of possible TEXTs depends on the CHAPTER: VOLUME.CHAPTER Possible TEXTs mv.01 01 02 03 ... 78 79 mv.02 01 02 03 ... 35 36 mv.03 01 02 03 ... 13 14 mv.04 01 02 03 ... 17 18 mv.05 01 02 03 ... 12 13 mv.06 01 02 03 ... 39 40 mv.07 01 02 03 ... 12 13 mv.08 01 02 03 ... 31 32 mv.09 01 02 03 ... 06 07 mv.10 01 02 03 04 05 06
SECTION:	The range of possible SECTIONs varies from TEXT to TEXT: 01 02 03 ... . TEXTs for which Horner has not enumerated any SECTIONs are assigned a SECTION of 01.
Examples:	mv.08.26.01-08.than.html (Mv 8.26.1-8 {Horner, Part IV, p.431}; the story of the monk with dysentery)

Cullavagga (Cv)

Note: ATI follows IB Horner's PTS numbering (Book of the Discipline, Part 5).

Form:	VOLUME.CHAPTER.TEXT.SECTION.SOURCE.EXT
VOLUME:	cv
CHAPTER:	01 02 03 ... 11 12
TEXT:	The range of possible TEXTs depends on the CHAPTER: VOLUME.CHAPTER Possible TEXTs cv.01 01 02 03 ... 33 34 cv.02 01 02 03 ... 08 09 cv.03 01 02 03 ... 35 36 cv.04 01 02 03 ... 13 14 cv.05 01 02 03 ... 36 37 cv.06 01 02 03 ... 20 21 cv.07 01 02 03 04 05 cv.08 01 02 03 ... 13 14 cv.09 01 02 03 04 05 cv.10 01 02 03 ... 26 27 cv.11 01 cv.12 01 02
SECTION:	The range of possible SECTIONs varies from TEXT to TEXT: 01 02 03 ... . TEXTs for which Horner has not enumerated any SECTIONs are assigned a SECTION of 01.
Examples:	cv.05.06.01x.olen.html (Cv 5.6, excerpt. Horner does not give any section numbers in Cv 5.6, so we call it 01.)

Digha Nikaya (DN)

Form:	VOLUME.TEXT.SECTION.SOURCE.EXT
VOLUME:	dn
TEXT:	01 02 03 ... 33 34
SECTION:	The range of possible SECTIONs depends on the TEXT: VOLUME.TEXT SECTION dn.01 1 2 3 dn.02 0 dn.03 1 2 dn.04 0 ... 0 dn.09 0 dn.10 1 2 dn.11 0 dn.12 0 dn.13 0 dn.14 1 2 3 dn.15 0 dn.16 1 2 3 4 5 6 dn.17 1 2 dn.18 0 dn.19 0 dn.20 0 dn.21 1 2 dn.22 0 dn.23 0 dn.24 1 2 dn.25 0 ... 0 dn.29 0 dn.30 1 2 dn.31 0 dn.32 0 dn.33 1 2 3 dn.34 1 2
Examples:	dn.01.2.abcd.html (DN 1, section 2) dn.16.5-6.than.html (DN 16, sections 5-6, complete) dn.16.1-6.vaji.html (DN 16, complete (i.e., sections 5-6)) dn.16.1-3x.abcd.html (DN 16, sections 1-3, excerpts) dn.22.0.than.html (DN 22, complete)

Majjhima Nikaya (MN)

Form:	VOLUME.TEXT.SOURCE.EXT
VOLUME:	mn
TEXT:	001 002 003 ... 151 152
Examples:	mn.001.than.html (MN 1, complete) mn.021x.budd.html (MN 21, excerpt)

Samyutta Nikaya (SN) 3

Form:	VOLUME.TEXT.SOURCE.EXT
VOLUME:	sn01 sn02 sn03 ... sn55 sn56
TEXT:	001 002 003 ...
Examples:	sn01.001.than.html (SN 1.1) sn36.010.nypo.html (SN 36.10) sn56.011.piya.html (SN 56.11)

Anguttara Nikaya (AN)

Form:	VOLUME.TEXT[.SECTION].SOURCE.EXT
VOLUME:	an01 an02 an03 ... an10 an11
TEXT:	001 002 003 ...
SECTION:	01 02 03 ... A few suttas contain enumerated sections, which in other editions of the Tipitaka are treated as separate suttas. For example, AN 3.100 (PTS) contains 15 sections; in the Thai edition the first 10 are counted as one sutta and the last 5 as another. To avoid ambiguity, we therefore number the first as an03.100.01-10 and the second as an03.100.11-15.
Examples:	an01.021-040.than.html (AN 1.21 through AN 1.40, complete) an01.031-040x.wood.html (AN 1.31 through AN 1.40, excerpts) an03.100.01-10.than.html (AN 3.100, first part; see above re: SECTION) an05.161.nymo.html (AN 5.61)

Khuddakapatha (Khp)

Form:	VOLUME.TEXT.SOURCE.EXT
VOLUME:	khp
TEXT:	1 2 3 ... 8 9
Examples:	khp.9.amar.html (Khp 9) khp.1-9x.piya.html (Khp 1-9, excerpts)

Dhammapada (Dhp)

Form:	VOLUME.TEXT.SOURCE.EXT
VOLUME:	dhp
TEXT:	01 02 03 ... 25 26
Examples:	dhp.06.than.html (Dhp, Pandita vagga) dhp.23.budd.html (Dhp, Naga vagga) dhp.14.than.html#dhp-183 (Dhp 183; see note)
Note:	Individual verses are referenced by href anchors, which are numbered consecutively across vaggas. Thus, Dhp 183 would be referenced as <a href="dhp.14.abcd.html#dhp-183">....

Udana (Ud)

Form:	VOLUME.CHAPTER.TEXT.SOURCE.EXT
VOLUME:	ud
CHAPTER:	1 2 3 4 5 6 7 8
TEXT:	01 02 03 ... 09 10
Examples:	ud.1.02.abcd.html (Ud 1.2)

Itivuttaka (Iti)

Form:	VOLUME.CHAPTER.TEXT.SOURCE.EXT
VOLUME:	iti
CHAPTER:	1 2 3 4
TEXT:	001 002 003 ... 111 112
Examples:	iti.1.001.abcd.html (Iti 1) iti.1.002.abcd.html (Iti 2) iti.2.028-030.abcd.html (Iti 28-30) iti.4.106-112x.irel.html (excerpts from Iti 106-112)
Note:	The numbering of suttas (TEXTs) runs consecutively across chapters (it does not restart at the beginning of each chapter).

Suttanipata (Sn)

Form:	VOLUME.CHAPTER.TEXT.SOURCE.EXT
VOLUME:	snp
CHAPTER:	1 2 3 4 5
TEXT:	01 02 03 ... 15 16
Examples:	snp.1.01.abcd.html (Sn 1.1) snp.1.02.abcd.html (Sn 1.2) snp.5.16.abcd.html (Sn 5.16)

Vimanavatthu (Vv)

Form:	VOLUME.CHAPTER.TEXT.SOURCE.EXT
VOLUME:	vv
CHAPTER:	1 2 3 4 5 6 7
TEXT:	01 02 03 ... 83 84 85
Example:	vv.1.16.irel.html (Vv 1.16)

Petavatthu (Pv)

Form:	VOLUME.CHAPTER.TEXT.SOURCE.EXT
VOLUME:	pv
CHAPTER:	1 2 3 4
TEXT:	01 02 03 ... 49 50 51
Example:	pv.1.05.than.html (Pv 1.5)

Theragatha (Thag)

Form:	VOLUME.CHAPTER.TEXT.SOURCE.EXT
VOLUME:	thag
CHAPTER:	01 02 03 ... 20 21
TEXT:	01 02 03 ... 48 49
Examples:	thag.01.00.abcd.html#sutta001 (Thag 1.1; see Note) thag.01.00.abcd.html#sutta120 (Thag 1.120; see Note) thag.02.01.abcd.html (Thag 2.1) thag.02.49x.abcd.html (Thag 2.49 (excerpt)) thag.21.01.abcd.html (Thag 21.1)
Note:	Chapter 1 contains 120 one-liner "suttas". Rather than put each of these in individual files, Chapter 1 is treated as a monolithic entity, with TEXT assigned the value 00 and individual verses referenced by href anchors. Chapter 2 has the next-highest number of verses (49), so 49 is the maximum possible TEXT number. Some chapters (e.g., 21) contain only one sutta, which is numbered 1.

Therigatha (Thig)

Form:	VOLUME.CHAPTER.TEXT.SOURCE.EXT
VOLUME:	thig
CHAPTER:	01 02 03 ... 15 16
TEXT:	01 02 03 ... 09 10
Examples:	thig.01.00.abcd.html#sutta01 (Thig 1.1; see Note) thig.01.00.abcd.html#sutta18 (Thig 1.18; see Note) thig.02.03.abcd.html (Thig 2.3) thig.12.01x.abcd.html (Thig 12.1 (excerpt))
Note:	Chapter 1 contains 18 one-liner "suttas". Rather than put each of these in individual files, Chapter 1 is treated as a monolithic entity, with TEXT assigned the value 00 and individual verses referenced by href anchors. Chapter 2 has the next-highest number of verses (10), so 10 is the maximum possible TEXT number. Some chapters (e.g., 12) contain only one sutta, which is numbered 1.

Mahaniddesa (Nm)

Form:	VOLUME.CHAPTER.TEXT.SOURCE.EXT
VOLUME:	nm
CHAPTER:	1 2 3 ... 15 16
TEXT:	01 02 03 ... 20 21
Example:	nm.2.04.olen.html (Nm 2.4)
Note:	CHAPTER and TEXT are taken from the BJT edition.

Milindapañha (Miln)

Horner's strange enumeration of sections in Miln frustrates attempts at building logical file names. Suttas are grouped in "divisions," each of which contains several questions posed to Ven. Nagasena. Her chapter II contains divisions 1-3, while chapter III contains divisions 4-7. Thus, if the last question in chapter II were referenced as "Miln 2.3.16" (Horner p. 87), the next question, in chapter III, would be referenced as "Miln 3.4.1" (p. 89). In other words, there is no "Miln 3.1.1". So be it.

Form:	VOLUME.CHAPTER.[DIVISION.]TEXT.SOURCE.EXT
VOLUME:	miln
CHAPTER:	1 2 3 4 5 6 7
DIVISION:	depends on the CHAPTER.
TEXT:	dependson the DIVISION.
	VOLUME.CHAPTER DIVISION TEXT miln.1 - - [not numbered ] miln.2 1 01 02 ... 15 2 1 2 ... 9 3 01 02 ... 16 miln.3 4 1 2 ... 8 5 01 02 ... 10 6 01 02 ... 11 7 01 02 ... 16 miln.4 1 1 2 ... 8 2 1 2 ... 8 3 01 02 ... 12 4 01 02 ... 10 5 01 02 ... 11 6 1 2 ... 9 7 01 02 ... 10 8 01 02 ... 12 miln.5 - - [not numbered ] miln.6 - 1 2 ... 9 miln.7 - 01 02 ... 67
Example:	miln.5x.olen.html (Miln 5, excerpts) miln.2.3.12.kell.html (Miln 2.3.12; Horner vol 1 p.85; PTS p.62) miln.2x.kell.html (Miln 2, excerpts)

Reserved Translator Codes

These four-byte codes are used in the translator code sub-field of a filename.