Markup Languages ML

1
MarkupLanguagesML
SSS
A
legal-X
SG
VOX
CP
X
DHT
G
HT
math
W
DS

• Yaakov J. Stein
• Chief ScientistRAD Data Communications

C
2
What do I do?
I digest, edit and produce documents

• business letters
• email
• meeting summaries
• proposals
• reports
• requirement specifications
• project plans

• web pages
• research articles
• review articles
• books

3
What do others do?

• Pretty much the same
• US corporations produce gt100 billion documents
  per year
• 90 of a modern institutions information is in
  documents
• gt50 of typical corporations efforts involves
  documents
• Thats why word processing SW
• was expected to bring efficiency increases
• But didnt!

4
Word processing?

• PROs
• makes nicer looking documents
• expedites document sharing during creation
• CONs
• typically 30 of effort on format and reformat
• doesnt increase information accessibility
• doesnt facilitate information mining

5
Databases?

• The natural alternative to documents are
  databases
• PROs
• increase information accessibility
• facilitate information mining
• CONs
• not human readable
• format inflexible

6
The solution

• What we really want is to write unconstrained
  text
• but to have information retrieval as well !
• Method 1 Automatic text analysis
• AI program analyzes text
• Recognizes document structure, sentence syntax
• Performs gisting, facilitates information mining
• Complete solution equivalent to solving Turing
  test
• Method 2 Manual markup
• Document author responsible for marking
• Clarifies document structure
• Enables automated retrieval of selected
  information
• Suggests presentation format

7
Why is text analysis hard?

• The man cried FIRE !

The man cried FIRE the gun !
The man cried FIRE the gun maker !
8
Are MLs computer languages?

• There are many different types of computer
  languages
• procedural languages
• for (n0nlt10i)
• if (ngt5) printf(markup languages are fun!\n)
• graphic languages
• newpath
• 0 0 moveto 0 1 lineto 1 1 lineto 1 0 lineto
  
• closepath fill
• database languages
• SELECT book FROM biblio WHERE subjectDSP AND
  authorSTEIN
• logical languages
• useful(DSP), useful(hardware), fun(DSP), fun(web)
• interesting(X) if useful(X) and fun(X)
• ?-interesting(X)

9
They are!

• Markup languages do not directly instruct
  computers
• like procedural languages
• rather indirectly instruct computer
• like logical languages
• They do this by using
• elements
• attributes
• entities
• text

ltBOOK SUBJECTdspgt ltTITLE
FORMATshortgtDSP-CSPlt/TITLEgt ltAUTHORgtJ.
Steinlt/AUTHORgt This is a great book!
standard-disclaimer lt/BOOKgt
(tags)
10
Some markup element functions

• Structural
• Clarifies document structure
• Delineates document parts
• Descriptive (informative)
• Indicates
• Facilitates information retrieval
• Presentational (display)
• Presents information in nice format
• Helps human readability
• Referential (links, applications)
• Provide hypertext links
• Launch applications

11
Structural Markup

• ltHEADINGgtSeptember 1, 2000lt/HEADINGgt
• ltGREETINGgtDear Prof. Stein, lt/GREETINGgt
• ltBODYgt
• I would like to tell you how much I enjoyed
  reading your new text
• Digital Signal Processing, A Computer Science
  Perspective.
• I hope we will be able to meet at the next
  conference.
• lt/BODYgt
• ltSIGNATUREgt
• Sincerely,
• Dee Espy
• lt/SIGNATUREgt

12
Descriptive Markup

• ltDATEgtSeptember 1, 2000lt/DATEgt
• Dear ltPERSONgtProf. Stein,lt/PERSONgt
• I would like to tell you how much I enjoyed
  reading your new text
• ltBOOKgt
• Digital Signal Processing, A Computer Science
  Perspective.
• lt/BOOKgt
• I hope we will be able to meet at the next
  ltEVENTgtconference.lt/EVENTgt
• Sincerely,
• ltPERSONgtDee Espylt/PERSONgt

13
Presentational Markup

• ltRIGHT-JUSTIFYgtSeptember 1, 2000lt/RIGHT-JUSTIFYgt
• ltBOLDgtDear Prof. Stein,lt/BOLDgt
• I would like to tell you how much I enjoyed
  reading your new text
• ltUNDERLINEgt
• Digital Signal Processing, A Computer Science
  Perspective.
• lt/UNDERLINEgt
• I hope we will be able to meet at the next
• ltBLINKgtconference.lt/BLINKgt
• Sincerely,
• ltIMAGE SRCdeesignature.jpg ALIGNleftgt
• ltFONT FACETimes-RomangtDee Espylt/FONTgt

14
Relational Markup

• lttoday xlinkformsimple hrefdate
  actuateautogt
• Dear Prof. Stein,
• I would like to tell you how much I enjoyed
  reading your new text
• ltA HREFwww.amazon.com/exec/obidos/ASIN/04712954
  gt
• Digital Signal Processing, A Computer Science
  Perspective.
• lt/Agt
• I hope we will be able to meet at the next
• ltA HREFconferencegtconference.lt/Agt
• Sincerely,
• ltIMAGE SRCdee-signature.jpg ALIGNleftgt
• ltA HREFmailtodee_at_dee-epsy.netgtDee Espylt/Agt

15
Generalized Markup Language

• William Tunnicliffe, Stanley Rice 1960s
• (independently) invent idea of structural markup
  language
• Problem need different ML for each type of
  document
• (letter, report, article, book,
  etc)
• Charles Goldfarb, Edward Mosher, Raymond Lorie
  (IBM) 1973
• invent Generalized Markup Language (GML)
• Solution use metalanguage
• Document Type Definition (DTD)
  defines tags
• IBM marked up 90 of its documents with GML

16
With GML structure is evident

• Library
• Novels
• Journals
• Textbooks
• Algebraic zoology
• Botanical history
• Computer poetry
• DSP
• DSP-CSP
• DSP just for fun
• Elementary QED

• Title
• Full Digital Signal Processing
• a Computer Science Perspective
• Short DSPCSP
• Author
• Name Jonathan (Y) Stein
• Association RAD Data Comm.
• Publication
• Publisher John Wiley
• Year 2000
• Location New York
• ISBN 04712954

17
Standard Generalized Markup Language

• Problems with GML
• No validating parser
• Not portable (between computer systems)
• Solution
• SGML
• ANSI 1978
• ISO/IEC 8879 1986 (Intl Org for Standardization
  / Intl Electrotechnical Commission)
• JTC1/SC34/WG1 (WG 1 of SubCommittee 34 of Joint
  Technical Committee 1)
• For presentation
• Document Style Semantics and Specification
  Language

18
SGML - cont.

• If SGML is so good why doesnt anyone use it ?
• Complexity
• base standard gt500 pages
• SGML is a metalanguage
• writing DTD is complex programming
• marked up text is hard to read
• DSSSL adds to complexity
• Inflexibility - requires absolute conformity
• assumes only one correct way to markup
• constrains author to dictated structure
• not good at capturing authors structure

19
HyperText Markup Language

• CERN (particle physics institute in Switzerland)
  was an early Internet adopter
• Used extensively for collaboration (articles have
  long author lists)
• Major problems with format incompatibility
• only straight ASCII worked reliably
• Tim Berners-Lee (computer specialist) defined
  requirements
• simplicity (couldnt expect physicists to use
  SGML)
• freedom (didnt need validation, let browser
  ignore bad markup)
• needed hypertext links (including to documents
  over Internet)
• presentational markup (papers must look nice -
  authors used to TEX)
• Solution HTML - a specific application of SGML
  (not metalanguage)

20
HTML versions

• HTML 1.0 (1989) Berners-Lee original CERN version
• hypertext, images, headbody structure,
  presentational markup
• HTML 2.0 (1994) IETF standard - RFC 1866
• added lists, forms, etc.
• HTML 3.2 (1997) W3C recommendation (incorporates
  Netscape extensions)
• added tables, applets, super/sub-scripts
• HTML 4.0 (1997) W3C recommendation (and similar
  ISO/IEC 15445)
• minimizes presentational markup
• XHTML 1.0 (2000) present W3C recommendation
• reformulates HTML in XML

21
HTML document structure

• ltHTMLgt
• ltHEADgt
• global definitions such as
• ltTITLEgtWeb page titlelt/TITLEgt
• lt/HEADgt
• ltBODYgt
• marked-up text
• lt/BODYgt
• lt/HTMLgt

22
Some HTML (body) elements

• ltH1gtLevel 1 Headinglt/H1gt Level 1
  Heading
• ltH2gtLevel 2 Headinglt/H2gt Level 2
  Heading
• ltH3gtLevel 3 Headinglt/H3gt Level 3
  Heading
• ltEMgt emphasized lt/EMgt
  emphasized
• ltPgt Paragraph lt/Pgt
  Paragraph
• ltA HREFurlgtlinklt/Agt link
• ltULgt
• ltLIgt item 1 lt/LIgt
  .item 1
• ltLIgt item 2 lt/LIgt
  . item 2
• lt/ULgt
• ltOLgt
• ltLIgt item 1 lt/LIgt
  1 item 1
• ltLIgt item 2 lt/LIgt
  2 item 2
• lt/OLgt
• ltIMG SRCurlgt

23
Problems with HTML

• Presentational aspects have predominated
• ltBgt bold text lt/Bgt
• ltBLINKgt blinking text lt/BLINKgt
• ltFONT COLORredgt red text lt/FONTgt
• Practically no descriptive markup
• Search engines are reduced to flat text search
• Search by topic only through keywords or portals
• Not extensible
• Cant add new tags
• Unknown tags ignored
• Links are relatively simple
• Usually user action is required (except IMG)
• Only full document (with offset) linkable
• Link management is logistic nightmare

24
Not everything is HTML

• Due to HTML limitations other tools are also
  used
• Multimedia extensions
• (dynamic) gif, jpg,
• streaming audio
• Common Gateway Interface
• generate HTML on-the-fly
• Perl, C,
• Server Push - Server Pull
• Javascript
• Java

25
eXtensible Markup Language

• Simplified (best parts of) SGML (subset of
  features)
• Flexible content management tool
• W3C recommendation(s)
• Extensible - can add new elements (even without
  DTD)
• Easy to create special purpose languages (with
  DTD/SCHEMA)
• Includes HTML-like hypertext links
• and extensions (XLINK, XPOINTER)
• The future of the web !

26
XML - an Example

• lt?xml version"1.0" standalone"yes"?gt
• ltbibliographygt
• ltbook isbn04712954gt
• lttitlegtDigital Signal Processing a Computer
  Science Perspectivelt/titlegt
• ltauthorgtJonathan (Y) Steinlt/authorgt
• ltpublishergtJohn Wiley and Sonslt/publishergt
• lt/bookgt
• ltarticlegt
• lttitlegtFalse Alarm Reduction for ASR and
  OCRlt/titlegt
• ltauthorgtYaakov Steinlt/authorgt
• ltproceedingsgtTenth AICVNN Symposiumlt/proceedings
  gt
• ltpagesgt195-200lt/pagesgt
• lt/articlegt
• ...
• lt/bibliographygt

27
What can we do with an XML file?

• Check if well-formed
• Check if valid (against DTD or schema)
• Display as-is in browser
• Parse in special-purpose program (SAX, DOM)
• Process (XSL) to XML, HTML, etc.
• Display after processing

28
Wireless Markup Language

• Markup language element of Wireless Application
  Protocol
• WAP forum (1997)
• Ericsson, Motorola, Nokia, Unwired Planet
  (phone.com)
• bring Internet to cellular phone users
• re-use fundamental Internet concepts (TCP/IP,
  http, html, javascript)
• but adapted to lower bandwidth
• smaller screen
• limited input facilities
• limited computational resources
• applications scale across transport options (GSM,
  TDMA, CDMA, 3G)
• and device types (mobile phones, personal
  assistants)

29
WML Philosophy

• Defined using XML
• Transported in compressed binary (for BW
  reduction)
• Applications are modeled as decks of cards
• Features
• Actions (OK, navigation, help) can be performed
• Hyperlinks (like in HTML)
• String variables
• Timers
• wbmp images (BW)
• Select boxes, forms (for input)
• wmlscript (like javascript)

30
WML structure

• lt ? xml version1.0 ? gt
• lt!DOCTYPE wml gt
• ltwmlgt
• ltcardgt
• ltpgt
• text
• lt/pgt
• ltpgt
• text
• lt/pgt
• lt/cardgt
• ltcardgt
• ...
• lt/cardgt
• lt/wmlgt

31
Some WML elements

• ltpgt lt/pgt text
• lta href...gt lt/agt hyperlink (anchor)
• ltdogt lt/dogt action
• ltgo href.../gt goto wml page
• lttimergt trigger event
  (units tenths of a second)
• ltinput/gt input user text
• ltprev/gt return to previous
  page
• () value of variable
• ltimg src /gt display image
• ltpostfield name value/gt set
  variable
• ltselect gt ltoptiongt ltoptiongt lt/selectgt select
  box

32
Some more markup languages

• VML Vector (graphics) Markup Language
• VoiceXML
• SSML Speech Synthesis Markup Language
• CPML Call Policy Markup Language
• DSML Directory Services Markup Language
• MathML Mathematical Markup Language
• CML Chemical Markup Language
• AML Astronomical Markup Language
• LegalXML
• BSML Bioinformatic Sequence Markup Language
• GedML Genealogical Data Markup Language
• FinXML Financial market Markup Language
• ChessML
• SDML Signed Document Markup Language
• RELML Real Estate Listing Markup Language
• etc. etc. etc. ...

33
Examples

• HTML
• html examples
• XML
• xml-file xsl-file xml
• VML
• vml-file
• WML (get M3gate emulator)
• wml examples