Project for me

[Malcolm McLean]

Meeting with the boss the other day, and the upshot is that I've got
three months to do exactly as I like. It doesn't have to be anything of
any commercial value to the company. Part of the idea is to give me a
bit of a rest, so nothing too ambitious.

Anyone any suggestions?

--
Check out Basic Algorithms and my other books:
https://www.lulu.com/spotlight/bgy1mm

[Janis Papanagnou]

On 18.02.2024 14:33, Malcolm McLean wrote:
> Meeting with the boss the other day, and the upshot is that I've got
> three months to do exactly as I like. It doesn't have to be anything of
> any commercial value to the company. Part of the idea is to give me a
> bit of a rest, so nothing too ambitious.

That's a fine move. And three months an amazing time frame.

> Anyone any suggestions?

I would nonetheless do something of value for the company (but not
necessarily of "commercial value").

But without knowing your company it's hard to tell what exactly.

I'd inspect where there's some badly working processes that could
be enhanced by tool support. Or something to make development more
reliable. Or get and combine statistical data that makes it possible
to draw valuable conclusions for the domain you're working in, and
that can maybe also be used for (internal or external) PR activities.
Some tool or system to enhance QA. Or just a recreational tool, a
game, or some such, maybe. Or examine how to use existing free tools
to your (or your company's) benefit, and if necessary programming
adapters/enhancements to be able to use them. Contributing to any
free tool that you think is worth fostering. - So far just a few
random thoughts. - Depending on your choice it might not be a rest.

Janis

[songbird]

Malcolm McLean wrote:
> Meeting with the boss the other day, and the upshot is that I've got
> three months to do exactly as I like. It doesn't have to be anything of
> any commercial value to the company. Part of the idea is to give me a
> bit of a rest, so nothing too ambitious.

do you have to still be there? does it have to be
computer related?



> Anyone any suggestions?

a volunteer project doing something physical.

going through the phone answering system and making sure
it is functional and useful (and has non-annoying music or
...)...


songbird

[Malcolm McLean]

On 18/02/2024 16:19, songbird wrote:
> Malcolm McLean wrote:
>> Meeting with the boss the other day, and the upshot is that I've got
>> three months to do exactly as I like. It doesn't have to be anything of
>> any commercial value to the company. Part of the idea is to give me a
>> bit of a rest, so nothing too ambitious.
>
> do you have to still be there? does it have to be
> computer related?
>

The marketing is done from a physical office. Except for some of the web
related work, the software engineering is virtual and we have people
widely scattered throughout the world. I work from home. I can do
anything I want and it doesn't even have to be computer related. But
because I'm posting here of course I'm thinking mainly about computer
programming projects.

[Derek]

Malcolm,

> Anyone any suggestions?

If you fancy doing some software engineering research,
you could have a look at my book Evidence-based Software
Engineering for ideas that might catch your interest.

This book discusses what is currently known about software
engineering, based on an analysis of all the publicly available data.
pdf+code+all data freely available here
http://knosof.co.uk/ESEUR/

[Derek]

[Thiago Adams]

Very nice. I will read your book later.

I liked this part.

7.2 Desirable characteristics
What characteristics are desirable in source code?
This section assumes that desirable characteristics are those
that minimise one or more developer resources (e.g., cost,
time), and developer mistakes (figure 6.9 suggests that most
of the code containing mistakes is modified/deleted before a
fault is reported). Desirable characteristics include:

• developers’ primary interaction with source code is read-
ing it to obtain the information needed to get a job done.
The resources consumed by this interaction can be reduced
by:

– reducing the amount of code that needs to be read, or the
need to remember previously read code,

– increasing the memorability of the behavior of code re-
duces the cost of rereading it to obtain a good enough
understanding,

– reducing the cognitive effort needed to extract a good
enough understanding of the behavior of what has been
read,

– being consistent to support the use of existing informa-
tion foraging1491 skills, and reducing the need to remem-
ber exceptions,



In my view abstractions many times works against these desirable features.

[Derek]

Thiago,

> In my view abstractions many times works against these desirable features.

What is need is £5 million to run some experiments
https://shape-of-code.com/2022/03/20/study-of-developers-for-the-cost-of-a-phase-i-clinical-drug-trial/

[Paul Edwards]

On 18/02/24 21:33, Malcolm McLean wrote:

> Meeting with the boss the other day, and the upshot is that I've got
> three months to do exactly as I like. It doesn't have to be anything of
> any commercial value to the company. Part of the idea is to give me a
> bit of a rest, so nothing too ambitious.
>
> Anyone any suggestions?

Can you help with SubC?

I'm not good with complex algorithms, so there are
bugs that I can't fix.


E.g. I have this code:

D:\devel\subc\src>type foo.c
typedef int xxx;

int foo(void)
{
xxx yyy;

yyy = 4;
return (xxx)3;
}

D:\devel\subc\src>


and (with debug) it is giving this error:


D:\devel\subc\src>scc -S foo.c
hit here
p is 2
three
four
p is 7
three
error: foo.c: 8: void value in expression

D:\devel\subc\src>


Those "p" values are
2 - PINT
7 - PVOID


The code that is failing is here:

expr.c:

static node *exprlist(int *lv, int ckvoid) {
node *n, *n2;
int p;

n2 = NULL;
n = asgmnt(lv);
p = lv[LVPRIM];
printf("p is %d\n", p);
if (COMMA == Token) n = rvalue(n, lv);
while (COMMA == Token) {
Token = scan();
n2 = asgmnt(lv);
n2 = rvalue(n2, lv);
p = lv[LVPRIM];
printf("p is %d\n", p);
n = mkbinop(OP_COMMA, n, n2);
}
printf("three\n");
/* return not working here */
if (ckvoid) notvoid(p);
printf("four\n");
return n;
}

And at this point I don't know what to do.

I looked in decl.c and I thought this might be the right track:

void typedecl(void) {
int utype, prim;

Token = scan();
if (STRUCT == Token || UNION == Token) {
structdecl(CTYPE, UNION == Token);
}
else if (ENUM == Token) {
enumdecl(1);
}
else if ((utype = usertype(Text)) != 0) {
Token = scan();
decl(CTYPE, Prims[utype], utype);
}
else {
prim = primtype(Token, NULL);
printf("hit here\n");
Token = scan();
decl(CTYPE, prim, 0);

ie the decl call (below "hit here") may need to do something
more to set the primary type to int. ie that CTYPE needs to
be converted into a variable name in the "lv" array.

But it's beyond my current ability to solve.

If you can suggest something to try, I can
give that a go.


Note that if you are on the company time, I don't
want you to make any code changes, as they could
potentially claim copyright over it.

However, if you can simply give me advice, I can
try to make the code change myself.

The latest code is on the PDOS/386 disk (\devel\subc\src
in pdos.vhd in pdos.zip at http://pdos.org), but I can make
it available standalone or email if you are interested.

BFN. Paul.

[Chris M. Thomasson]

On 2/18/2024 5:33 AM, Malcolm McLean wrote:
> Meeting with the boss the other day, and the upshot is that I've got
> three months to do exactly as I like. It doesn't have to be anything of
> any commercial value to the company. Part of the idea is to give me a
> bit of a rest, so nothing too ambitious.
>
> Anyone any suggestions?
>

Create some kick ass graphics for fun. The math in all in... Fwiw, check
these out:

https://youtu.be/YsAkm0VlCsw
(Not to say this is kick ass, but it is 4k... ;^D)

https://youtu.be/XKhS_nklCkE
(my pseudo code in the comments... IFS, fun?)

Try to recreate my IFS?


https://youtu.be/n13GHyYEfLA
(make some music?)

https://youtu.be/DrPp6xfLe4Q
(fun with circles?)

https://youtu.be/hlHmOKvQVT4
(attack vectors? ;^))

[Chris M. Thomasson]

Btw, all in C++...