Spreadsheet
[1/14] from: andrew::wxc::net::nz at: 4-Mar-2001 10:17
Has anyone thought of making a spreadsheet in Rebol/View or Rebol/Link?
I'd really like to combine the best parts of Star Office Spreadsheet,
Microsoft Office Excel and Lotus Improv, and merge them together with a
generous helping of Rebol Script.
Andrew Martin
ICQ: 26227169 http://members.nbci.com/AndrewMartin/
[2/14] from: agem:crosswinds at: 3-Mar-2001 23:52
>>>>>>>>>>>>>>>>>> Ursprüngliche Nachricht <<<<<<<<<<<<<<<<<<
Am 03.03.01, 22:17:47, schrieb "Andrew Martin" <[andrew--wxc--net--nz]> zum
Thema [REBOL] Spreadsheet:
> Has anyone thought of making a spreadsheet in Rebol/View or
Rebol/Link?
> I'd really like to combine the best parts of Star Office Spreadsheet,
> Microsoft Office Excel and Lotus Improv, and merge them together with
a
> generous helping of Rebol Script.
> Andrew Martin
> ICQ: 26227169 http://members.nbci.com/AndrewMartin/
> -><-
like this? (well, more a joke :)
[REBOL [title: file: %rebsheet2.r date: 6-Nov-2000/22:12:57+1:00
autor: 'volker]
formulas: none
do rebsheet: [
file: %rebsheet2.r
as-always: 640x480
list-size: 600x380
label-size: min 640x15 list-size / 5
if not formulas [formulas: array/initial [25 5] ""
formulas/1: copy/deep [
{/do[save-me] {remove "/" => autosave (lots of!)}}
"$123" "1.5 * sheet/1/2"
{(pick pick sheet y + 0 x - 2) + (pick pick sheet y + 0 x
- 1)}
"reduce[y x]"
]
]
sheet: copy/deep formulas
clear-sheet: does [foreach y sheet [forall y [change y 0]]]
save-me: does [
save/header file compose [formulas: (reduce [formulas])
do rebsheet: (reduce [rebsheet])]
compose [title: file: (file) date: (now) autor: 'volker]
]
set-text: func [face string] [face/text: string face/line-list:
none
show face]
ed-x: ed-y: 1
myst: stylize [
sheet-label: text label-size with [y: x: none action: [
poke formulas/:ed-y ed-x copy edit/text
set-text edit copy formulas/:y/:x
ed-y: y ed-x: x show sheet-list
]]
]
view layout [size as-always
styles myst
edit: area min list-size 640x1 * 1x4 * 1x15 formulas/1/1
sheet-list: list list-size 30.30.30 [
across sheet-label sheet-label sheet-label sheet-label
sheet-label
]
supply [
y: face/y: count x: face/x: index
either all [ed-x = x ed-y = y] [face/color: 0.0.255]
[face/color: 30.30.30]
if all [1 = count 1 = index] [clear-sheet]
face/text: either count > length? formulas ["-"] [
if error? try [res: do formulas/:y/:x] [res: "???"]
poke sheet/:y x :res
mold :res
]
]
]
]
]
[3/14] from: brett:codeconscious at: 4-Mar-2001 11:24
The thought occurred to me. I actually did come across a situation where we
wanted to provide spreadsheet functionality but in a very constrained
manner. The reason being that the users were already using spreadsheets and
were very nervous about causing errors. So they wanted a custom program
instead, but they needed flexibility in how to define a calculation.
We never did that app, because we were going to get a "better bang for buck"
by doing other things first.
Anyways my thoughts about doing this with Rebol:
1) Need to build a grid based entry interface style. Once done, very useful
for other Rebol applications.
2) Need to code a graph sorting algorithm that gets as input the
dependencies between cell calculations.
3) Need to provide all the little user-interface niceties that people expect
from spreadsheet applications.
4) Aim to keep as much power of Rebol as possible accessible to user entered
parts of the spreadsheet.
I think (1) is the most reusable for other applications. (2) is an
interesting problem to solve.
(3) may actually be the most amount of work in the project - users expect a
lot!
but (4) is a good justification for a Rebol powered spreadsheet - Rebol
programming can be made visual...
Brett.
[4/14] from: robbo1mark:aol at: 3-Mar-2001 20:33
Well do it then!!!
Who or what is stoppin you?
It would be great if you could achieve something
outwith REBOL?
Mark Dickson
[5/14] from: dvydra2:yaho:o at: 3-Mar-2001 21:07
Mark,
This remark strikes me as rude.
David
--- [Robbo1Mark--aol--com] wrote:
> Well do it then!!!
> Who or what is stoppin you?
<<quoted lines omitted: 6>>
> [rebol-request--rebol--com] with "unsubscribe" in the
> subject, without the quotes.
=====
please reply to: [david--vydra--net]
[6/14] from: agem:crosswinds at: 4-Mar-2001 7:41
>>>>>>>>>>>>>>>>>> Ursprüngliche Nachricht <<<<<<<<<<<<<<<<<<
Am 04.03.01, 01:24:20, schrieb "Brett Handley" <[brett--codeconscious--com]>
zum Thema [REBOL] Re: Spreadsheet:
> The thought occurred to me. I actually did come across a situation
where we
> wanted to provide spreadsheet functionality but in a very constrained
> manner. The reason being that the users were already using
spreadsheets and
> were very nervous about causing errors. So they wanted a custom
program
> instead, but they needed flexibility in how to define a calculation.
> We never did that app, because we were going to get a "better bang for
buck"
> by doing other things first.
> Anyways my thoughts about doing this with Rebol:
> 1) Need to build a grid based entry interface style. Once done, very
useful
> for other Rebol applications.
> 2) Need to code a graph sorting algorithm that gets as input the
> dependencies between cell calculations.
How much calculatios would be needed?
Is a left-right/up-bottom complete calculation really to slow?
Because you can enter simply rebol-code in cells,
and then dependency-chcking is hard..
> 3) Need to provide all the little user-interface niceties that people
expect
as a seldom spreadsheet-user, what do they expect?
Copy/paste of areas/formulas?
Or would a rebol-loop accessing cells are enough?
> from spreadsheet applications.
> 4) Aim to keep as much power of Rebol as possible accessible to user
entered
> parts of the spreadsheet.
> I think (1) is the most reusable for other applications. (2) is an
> interesting problem to solve.
> (3) may actually be the most amount of work in the project - users
expect a
> lot!
> but (4) is a good justification for a Rebol powered spreadsheet -
Rebol
> programming can be made visual...
> Brett.
Volker
[7/14] from: andrew:wxc at: 4-Mar-2001 21:17
> > 2) Need to code a graph sorting algorithm that gets as input the
dependencies between cell calculations.
> How much calculations would be needed? Is a left-right/up-bottom complete
calculation really to slow? Because you can enter simply rebol-code in
cells, and then dependency-checking is hard...
Copying formulas from cell to cell is boring for me. I'd prefer the Lotus
Improv method where formulas are separate from cells, the formulas then have
the ability to be rearranged so as to change dependencies.
Andrew Martin
ICQ: 26227169 http://members.nbci.com/AndrewMartin/
[8/14] from: chaz:innocent at: 4-Mar-2001 2:34
:) One way of finding one's place in a group is to collide full force into
every identifiable object until one discovers where the boundaries are.
When the maintainer of http://www.codeconscious.com/ invites people to
explore some interesting opportunities, it's kind of humorous to see him
encouraged to "do something outwith REBOL"
You've got a lot of energy.
And yes. The perl community is a beautiful and powerful thing, and perhaps
the potential for the REBOL community to grow into something like that
excites you.
And yes. Open-source software does promise to revolutionize the economics
of intellectual property, and perhaps the potential for REBOL to transform
economics appeals to you as well.
However, I have confidence that in the future, more and more diverse
communities will be saying, "We should be following the example of REBOL"
one example: http://www.cs.unlv.edu/~slumos/uniconmsg.html
chaz
At 08:33 PM 3/3/01 EST, you wrote:
[9/14] from: brett:codeconscious at: 4-Mar-2001 22:49
Thanks for replying Volker,
> How much calculatios would be needed?
The particular example requirement I described doesn't exist anymore - they
sold the company before I did it :)
>> 3) Need to provide all the little user-interface niceties that people
>> expect
> as a seldom spreadsheet-user, what do they expect?
> Copy/paste of areas/formulas?
> Or would a rebol-loop accessing cells are enough?
My comment was based on the situation where we were going to give our app as
a replacement for using MS Excel - for that particular need only of course.
So what do people expect? -> All the normal Windows behaviour, and Excel
behaviour otherwise there is frustration. It all depends on your audience
and how much training your are prepared to give them.
I'll now leave the specs for Reb-spread or whatever it is to Andrew - his
idea ;)
Brett.
[10/14] from: brett:codeconscious at: 4-Mar-2001 22:56
Hi Andrew,
> > > 2) Need to code a graph sorting algorithm that gets as input the
> dependencies between cell calculations.
>
> > How much calculations would be needed? Is a left-right/up-bottom
complete
> calculation really to slow? Because you can enter simply rebol-code in
> cells, and then dependency-checking is hard...
>
> Copying formulas from cell to cell is boring for me. I'd prefer the Lotus
> Improv method where formulas are separate from cells, the formulas then
have
> the ability to be rearranged so as to change dependencies.
I haven't used Improv. How are the cell contents "bound" to the formulas -
by a cell grid reference (like excel) or by a giving each cell a name, or by
pointing with cursor?
And the flip side, how is the output cell connected with the formula, by
entering a formula name, or what?
Curious.
Brett.
[11/14] from: robbo1mark:aol at: 4-Mar-2001 7:17
Sorry if my previous remarks came across a rude, this was not my intention
and I apologize here unreservedly for any offense caused.
I simply meant to make the point that if it is important enough to you then
why don't you go ahead and DO it, and why not try to enlist the help of
others if it is too big a project to manage on your own.
That's all, sorry for confusion or offense.
Mark Dickson
[12/14] from: agem:crosswinds at: 4-Mar-2001 12:36
>>>>>>>>>>>>>>>>>> Ursprüngliche Nachricht <<<<<<<<<<<<<<<<<<
Am 04.03.01, 09:17:15, schrieb "Andrew Martin" <[andrew--wxc--net--nz]> zum
Thema [REBOL] Re: Spreadsheet:
> > > 2) Need to code a graph sorting algorithm that gets as input the
> dependencies between cell calculations.
> > How much calculations would be needed? Is a left-right/up-bottom
complete
> calculation really to slow? Because you can enter simply rebol-code in
> cells, and then dependency-checking is hard...
> Copying formulas from cell to cell is boring for me. I'd prefer the
Lotus
> Improv method where formulas are separate from cells, the formulas
then have
> the ability to be rearranged so as to change dependencies.
Don't know this. How are cells and formulas are connected?
What do you think about my %rebheet3.r on the Volker-rebpage?
(function, not q&d-coding :)
[13/14] from: brett:codeconscious at: 10-Mar-2001 0:08
Hi Volker,
I'm sorry about taking a while to reply.
> What do you think about my %rebheet3.r on the Volker-rebpage?
> (function, not q&d-coding :)
I think it is great! You certainly have put a lot of work into it.
Implementing your own dialects for extra spreadsheet niceties was a great
idea.
I had fun playing with it. Putting in rebol expressions into one cell and
then having another reference it worked really well.
For example I read the contents of a directory into one cell, the next cell
iterated over each file in the block contained in the first cell and did
some silly processing.
The only thing was I couldn't see the result when it was too long -
otherwise it could be useful now.
It shows great promis don't you think?
It certainly gives ideas.
Brett.
[14/14] from: al:bri:xtra at: 11-Mar-2001 20:44
Volker wrote:
> What do you think about my %rebheet3.r on the Volker-rebpage? (function,
not q&d-coding :)
In case this was addressed to me, sorry, I haven't been feeling well, and
haven't yet made time available to check this out yet.
Andrew Martin
ICQ: 26227169 http://members.nbci.com/AndrewMartin/
Notes
- Quoted lines have been omitted from some messages.
View the message alone to see the lines that have been omitted