djh@metaphor.metaphor.com (Dallas Hodgson) (07/14/90)
In article <544@beguine.UUCP> Scott.Maxwell@samba.acs.unc.edu (BBS Account) writes: > > Uh, I'm a little confused by what I've read regarding Lattice C v. Aztec >C. Everybody seems to say (a) Aztec usually produces faster code, (b) Aztec >usually produces the code faster (that is to say, it compiles faster), >(c) Aztec is cheaper, and (d) Lattice is more widely used. > Now, that can't all be right -- unless Lattice has *vastly* superior >advertising or something, and I can't imagine that would be enough to make the >difference. Assuming I'm right about points (a), (b) and (c), though, why the >heck is (d) true? Does Lattice offer better support? What have I missed? Amiga commissioned Lattice to write the first commercially available Amiga compiler back in '85. This was to lend credibility to the machine, as Lattice was the leading C company for the PC at the time. (this is no longer the case!) For in-house development, Amiga was using the Greenhills compiler on the Sun which is probably still the best compiler available. Many developers were doing development on AT's using Lattice's 1.0 cross compiler and downloading the result to the Amiga over the parallel port, as there were no Amiga hard disks at that time. As far as I'm aware, Lattice's cross-compiler has never been updated. Jim Goodnow had beta copies of the first Manx compiler available for purchase at the 1986 Amiga DevCon 2 in Monterey. I bought my copy at that time. Most developers flamed Lattice endlessly until 5.0 came out, because Aztec always had the fastest compile time, produced much smaller executables, and supported precompiled headers. It was over a year after Lattice release 5.0 until Aztec released theirs. Folks were grumbling about the lack of support, and a lot of people switched over to Lattice. Both compilers are pretty similar now, although some folks might enjoy Manx's convenient #asm directives over Lattice's chip_mem keywords. I've tried both, and still use Manx. No matter what Lattice does, Manx still has the feel of compiling "tighter, smaller and faster". Jim Goodnow recently left Manx, so who's to say what future support from Manx will be. >Scott Maxwell (CSMAXWEL@ECUVM1.BITNET or Scott.Maxwell@samba.acs.unc.edu)
mwm@raven.pa.dec.com (Mike (Real Amigas have keyboard garages) Meyer) (07/18/90)
In article <1312@metaphor.Metaphor.COM> djh@metaphor.metaphor.com (Dallas Hodgson) writes:
Jim Goodnow recently left Manx, so who's to say what future support from
Manx will be.
In the process of trying to find a speaker for the next BADGE meeting,
I called Manx to see if they would be willing to address this subject
at the meeting (answer: no. - anyone wanna come give a talk?). They
pointed out that Jim had done less than 1/3rd of the work on 5.0, so
the change isn't as major as it looks from the outside. Further, Jim
would still be working with them on the 68000 product. He wasn't gone,
the relationship had just changed.
Back on the topic, I prefer Lattice because since at least 4.0, it
seems to catch more compile-time errors and have more informative
messages. From my experiences with Manx and watching the net, Manx has
more bugs.
Lattice has also always had excellent support, which doesn't seem to
be true of Manx. I don't expect this to change with SAS. SAS is at
least acting like they'll send someone to BADGE. It may not pan out,
but we'll have to see....
<mike
--
Come rain, come hail, come sleet, come snow Mike Meyer
You don't like to drive to slow mwm@relay.pa.dec.com
You're always in the passing lane decwrl!mwm
It's enough to drive me out of my brain