## In light of Jeff's passing...

forth

### Next

• 1. question about SSA algorithm
Suppose we have a control flow graph: 2 --> 7, 11 11 --> 13, 16 13 --> 17 16 --> 17 7 --> 17 And the dominance frontier for node 7 is 17 because there is an outgoing edge between 7 and 17 and 7 does not dominate 17 (2 does). Under SSA, if we have a variable that is defined at node 7, we will need to insert a phi function at node 17. Suppose there is a variable definition for i at instr 8, which is in node 7 of the CFG. So, we insert a phi for variable i at node 17. However, what if variable i has only one definition that is live at node 17? The other predecessors of 17, namely 13 and 16 DO NOT define i. This means that we can only put in 1 operand for phi as opposed to 3. However, the SSA algorithm says that phi will have 3 operands at node 17, because 17 has 3 ancestors. Please let me know if you know how to handle this case. Thanks, Shirley
• 2. Gforth: is new release ever to happen?
Hello! Anton, Bernd, I ask you once again, release! Is it so hard to make a tarball and label it "0.7.0"? You have enough numbers ahead, before you release 1.0, which I'm almost sure never to happen. Or publish any particular minor issues to resolve, if there're any. Where did my patch to support non-gnu make went? I haven't received any word in reply. Do you accept my mail at all? -- HE CE3OH...

### In light of Jeff's passing...

```It's surprising how much Jeff's death affected me this Wednesday.
Maybe it's because I'm getting older myself and was reminded of my own
mortality, being in my 40's and all..I didn't know Jeff personally but
I enjoyed reading his posts on c.l.f. and digging through
ultratechnology from time to time. Besides Chuck et al, he seemed to
be one of the few folks here who truly had embraced the forth
principle.

So that got me thinking...

There's an awful lot of knowledge that Jeff acquired (and contributed
to) by working for years with Chuck, Greg and the other folks at GA,
Intellasys, iTV, etc. Although he shared much of this online, his
personal experiences and the wealth of accumulated knowledge are now
gone forever.

I'd hate to see the details of Chuck's life-long pursuit of simpler,
faster technology be relegated to the dust-bin of the internet or
merely be a footnote in (presumably digital) books in 10-20 years.
Sure, there are interviews about Forth and all that, but there's
chip designs from an implementation point of view. I'd love to find
out more about that.. I'm sure Chuck has more important things to do
than write down his memoirs and experiences but hope that in time,
he'll find a way to share them with us.

Last but not least, I hope this post didn't offend anyone as that
wasn't my intention. FWIW, Jeff and his point of view will be sorely
missed on c.l.f.

-Mux
```

### Re: In light of Jeff's passing...

```In my opinion, that was always the problem with the work of Jeff,
Chuck, et al, they wanted to introduce a new technology, but they all
wanted to be millionares too. In effect Forth chips are a proprietary
technology, at least in their peak form. What if Chuck had instead
would be a commodity at this point offered by several manufatures,
like RAM. And Chuck would have made a good living as consultant too.
It seems like a good possibility to me, but I don't know exactly the
politics of the semi-conductor industry.

In the long term, Forth and probably stack machines will be confined
to the dustbin and will be a footnote in the history of computing.
require a different language and different algorithms. But digital
computing will be a around for a while, and the two stack architecture
is the best for factoring, passing parameters between nested
subroutines. Its a shame that the value of factoring and stack
machines has not really caught on. The RISC machines with the huge
register files don't benefit from factoring, you have to unfactor
```

### Re: In light of Jeff's passing...

```

how could you make a cheap commodity chip with a basic usible desktop
to repalce linux

once someone has that they can read web pages on howto program the
machien further

without that you tend to drift back to windows

I am a unix admin and love archlinux and icewm desktop and firefox

I use 1% or less of each.

if forth could offer cheap forth chip pc and broswer...then people
would get it and then perhaps learn forth as they go ocne they have
nice browsing platform.

I think this is way to forth success.
```

### Re: In light of Jeff's passing...

```

you mean jeff fox?

that really sucks

what happened?

amazing he did not offer a line fo lwo cost niternet pcs without 1
line of c

I would have purchased just to see how the alternative tek worked

esp if only few 100 \$
```

```It's surprising how much Jeff's death affected me this Wednesday.
Maybe it's because I'm getting older myself and was reminded of my own
mortality, being in my 40's and all..I didn't know Jeff personally but
I enjoyed reading his posts on c.l.f. and digging through
ultratechnology from time to time. Besides Chuck et al, he seemed to
be one of the few folks here who truly had embraced the forth
principle.

So that got me thinking...

There's an awful lot of knowledge that Jeff acquired (and contributed
to) by working for years with Chuck, Greg and the other folks at GA,
Intellasys, iTV, etc. Although he shared much of this online, his
personal experiences and the wealth of accumulated knowledge are now
gone forever.

I'd hate to see the details of Chuck's life-long pursuit of simpler,
faster technology be relegated to the dust-bin of the internet or
merely be a footnote in (presumably digital) books in 10-20 years.
Sure, there are interviews about Forth and all that, but there's
chip designs from an implementation point of view. I'd love to find
out more about that.. I'm sure Chuck has more important things to do
than write down his memoirs and experiences but hope that in time,
he'll find a way to share them with us.

Last but not least, I hope this post didn't offend anyone as that
wasn't my intention. FWIW, Jeff and his point of view will be sorely
missed on c.l.f.

-Mux
```

```Jeff Fox please write a book on learning forth the Jeff Fox way
```

```Hello Jeff,

I have a 1 line change suggestion for tcllib:nntp.tcl v1.7 but do not
know where to send it.

Just past line 125 the socket is opened.  Add a fconfigure to allow
the correct transmission
of binary (yEnc) messges.

set sock [socket \$data(host) \$data(port)]
fconfigure \$sock -encoding binary

Thank you for any help in passing this along.
```

```Your perception of this ng is accurate.  To clarify further, I generally
say that the content is:

65%: The endless topicality wars, with the featured attraction
being the frequent Heathfield/Navia battles.  This is obviously
the substance of the group, and what keeps people coming back
for more.

30%: Newbie questions, technically not off-topic, but often
bordering on trolling - given that they are often of the:
What does "i = i++ + ++i" do?
variety.  Note that even though these are not technically OT,
they are all covered in the available body of text which newbies
are expected to have read and digested before posting - i.e.,
the FAQs and the standards documents.  They are, by the
standards of most normal newsgroups, in the category of "bad
questions" - questions that should not be asked, since they've
all been asked many times before, and the answers have all been
written down.

5%: The remaining bits are the "language lawyering" stuff - of
great interest to those who care about such things, but utterly
boring to normal people like us.

So, there you have it.  HTH.

```