Path: utzoo!censor!geac!torsqnt!news-server.csri.toronto.edu!rutgers!usc!zaphod.mps.ohio-state.edu!think.com!mintaka!spdcc!esegue!compilers-sender
From: bright@nazgul.UUCP (Walter Bright)
Newsgroups: comp.compilers
Subject: Re: time to write a compiler
Keywords: C, question
Message-ID: <147@nazgul.UUCP>
Date: 8 Nov 90 02:36:00 GMT
References: <1990Oct31.180632.3201@nntp-server.caltech.edu>
Sender: compilers-sender@iecc.cambridge.ma.us
Reply-To: bright@nazgul.UUCP (Walter Bright)
Organization: Zortech, Seattle
Lines: 20
Approved: compilers@iecc.cambridge.ma.us

In article <1990Oct31.180632.3201@nntp-server.caltech.edu> roman@gaudi.ccsf.caltech.edu (Roman Salvador) writes:
<   I would like to find out the time it takes (more or less) to write a
<compiler (i.e. a Fortran 90 one). ...

It pretty much depends on if you are writing a research toy or plan to
sell the compiler. If you plan on selling it, triple all times you expect.

From personal experience, I started writing a C compiler in 1982. 2 years
later, it was shipped (Datalight C). I have been working on it steadilly
ever since. It will never be finished.

A global optimizer takes 1 year to write, and then 1 year to debug out in
the field. Some companies never get it to work reliably :-).

A C compiler is about twice as complex as a Pascal compiler. C++ is twice
as complex as C. Starting from a working C compiler, it takes 1 to 2 years
to add C++ capability to it.
-- 
Send compilers articles to compilers@iecc.cambridge.ma.us or
{ima | spdcc | world}!esegue!compilers.  Meta-mail to compilers-request.