Transcript AI 양식 하얀색 바탕

Standardization, Internationalization
Programming Language Design and Implementation (4th Edition)
by T. Pratt and M. Zelkowitz
Prentice Hall, 2001
Section 1.3.3-1.3.4
Who defines a language?

Is: I = 1 && 2 + 3 | 4; legal in C? What is assigned
to I if it is?


3 ways typically to answer this:
1. Read language manual (Problem: Can you find one?)
 부정확
2. Read language standard (Problem: Have you ever
seen it?)  BNF, Vienna definition language,
3. Write a program to see what happens. (Easy to do!)
 C의 표준은 CC (UNIX)



Most do 3, but current compilers may not give correct
answer
2
Creation of standards






Language standards defined by national standards bodies:
ISO - International Standards organization
IEEE - Institute of Electrical and Electronics Engineers
ANSI - American National Standards Institute
 code, 언어, 통신규약, XML, 전자책, …
WWW Consortium, EPC Global, …

SMTP(7bits)  MIME(8bits) : Plain text, 음성, 영상, ..





All work in a similar way:
1. Working group of volunteers set up to define standard
2. Agree on features for new standard
3. Vote on standard
4. If approved by working group, submitted to parent
organization for approval.
3
Creation of standards

Standards in the US are voluntary:


There is no federal standards-making organization.
NIST - National Institute for Standards and
Technology develops standards that are only required
on federal agencies, not for commercial organizations.

But: Consensus is the key to standards making:



Contentious features often omitted to gain consensus
Only vendors have a vested interest in the results
Users don't care until standard approved, and then it
is too late!
한국 ::: 표준연구소, 지경부, 교과부

4
Standards conforming programs

Standards define behavior for a standards conforming
program - one that meets the rules of the language standard

In general
program is
conforming
though the
–
(except for Ada), behavior of non-conforming
not specified, so any extensions to a standards
compiler may still be standards conforming, even
program is not standards conforming.
Java와 관련된 MS 전략

Standards supposed to be reviewed every 5 years

Examples:
–
–

FORTRAN 1966, 1977, 1990, 1995, 2003
Ada 1983, 1995
Not quite 5 years, but at least periodically
5
표준과 특허
 De facto 표준
 ISO에서는 주로 특허로 제한된 기술은
표준으로 정하지 않음
 국제표준에 따라 정해진 내용이 특허로
되면 심각한 법적 문제를 낳을 수 있음
 특허에 의해 “ㅈㅈㅈ” -> “www”로 고치
지 못함 ::: 특허와 관련된 몇 사례!!!
6
숙제

한글코드의 표준화 과정을 조사한다. 특히 Unicode가 어떻게 표준화되었
는지 조사한다. 또 KS5601, ISo646, ISO2422, UTF8, UTF16은 무엇인가?
– 정희성 박사 특허



자판을 중심으로 표준화가 낳을 수 있는 문제점을 조사한다. 실제 우리가
쓰는 자판은 최대한 글쇠가 엉키지 않게 만듦으로써 오히려 입력속도가 느
리다(QWERT). 이에 따라 새로운 자판(이름을 조사)이 나왔지만 익숙하지
않아서 아직 보급되지 않고 있다. 한글자판도 2벌식과 3벌식이 있는데 비
슷한 문제점이 있다.
현재 XML을 이용하여 각 분야의 문서표준화가 진행되고 있다. 그 예와 현
재 진행상황을 설명하라!! (예, eBook, vXML, 전자상거래 …), 한 분야만
잡아서 조사한다.
핸드폰과 스마트폰에서 한글 문자를 입력하는 방법은 표준화되어 있는가?
현재 표준화 상황을 조사하라.
 C, C++, Pyton, PHP, HTML, XHTML의 표준화 과정을 설명하라.
7
When to standardize a language?

Problem: When to standardize a language?
– If too late - many incompatible versions - FORTRAN in
1960s was already a de facto standard, but no two were
the same
– If too early - no experience with language - Ada in 1983
had no running compilers
– Just right - Probably Pascal in 1983, although it is
rapidly becoming a dead language

Other languages:
–
–
–
–
–
C in 1988 : ANSI C
LISP in 1990 - Way too late : common LISP
De facto standards: ML - One major implementation - SML
Smalltalk - none
Prolog - none
8
Internationalization

I18N issue - Internationalization - How to specify
languages useful in a global economy?

Characters used internationally:
– Single 8-bit byte; usual format today - 256 character
values. A lot in 1963, but insufficient today
– ASCII is a 7 bit 128 character code

What about other languages?
– Additional letters: German umlaut-ä, French accent-é,
Scandanavian symbols-ö,
– Russian, other alphabets (Greek, Arabic, Hebrew),
ideographs (Chinese), 한글
– Unicode - 16 bit code allows for 65K symbols. 8-bit byte
is insufficient
9
Internationalization (continued)

I18N name avoids deciding between
internationalization and localization



Some of the internationalization issues:
What character codes to use?
Collating sequences? - How do you alphabetize various
languages?
Dates? - What date is 10/12/01? Is it a date in
October or December?
Time? - How do you handle time zones, summer time in
Europe, daylight savings time in US, Southern
hemisphere is 6 months out of phase with northern
hemisphere, Date to change from summer to standard
time is not consistent.
Currency? - How to handle dollars, pounds, marks,
francs, euros, 원, etc.
 XML에서는 dublin core를 사용




10
Summary

Language design today must:



Allow program solution to match physical structure of
problem
Allow for world-wide use
Be easy to prove solution correct

Rest of course will work on these goals
11
Programming Environment


The environments in which programs are created and
tested
Separate compilation
– Co-development of any large program
– Function prototypes, external variables(COMMON,
extern), naming convention(_…), Scope rule (Pascal,
C, Ada), inheritance (Ada, C++), polymorphism,
overloaded (Ada)
– Stub : a subprogram call made to a subprogram that
has not yet been compiled
– To provide information for separate compilation
• FOTRAN : COMMON <- redeclaration
• Compile 순서 관리 <- Ada.
• A library containing specifications
– Consistency of the external data or subprograms
• Linking?
12
Programming Environment (cont.)


Testing and debugging
– Execution trace features (LISP, Prolog, …
debugger)
• Statements, variables
– Breakpoints
– Assertions
• Assert(X>0)
C
% Exception handling
13