Quote

Is this something that can be done within the builders' wiki

Well at least I guess so..
mediawiki has got some hooks we can step into to hack in a "userschoice" function that may work.
I cannot tell exactly because I never looked at the sourcecodes of mediawiki close enough to verify, but I think with some effort it's possible.. so yes

At least it is possible to add scripts to a wiki So we can go the cookie and javascript path if there is no other chance to do it.
I think I've seen some ajax extension too, that be pretty much perfect for this purpose if we can't use plain php.

I'll go and install mediawiki on my serever tomorrow to see what is possible and what is not, maybe it's a matter of minutes to get it done the way I mentioned

'sid

I've thrown together a somewhat prescriptive page

[objects.reprap.org]

It is far from complete, but I think it's quite easy to follow what is needed from it. I've mostly been comming from the building angle so there isn't anything about software versions.

Steve: I don't see why linux would be the prefered environment for the easy route or is it root ;-}

Here's my take...

degroof Wrote:
-------------------------------------------------------
>
> Just to throw something out there as a "standard"
> build, how about:
> - Arduino electronics

good.

> - Linux platform with side notes for Mac and
> Windows

has anybody done a survey to discover what percentage of users are using which OS. I'd say that you ought to have this choice focus on your biggest demographic. I've had to say over and over here, too often to a deaf audience, that the point of Reprap is to get Reprap machines running out there, not to shill for Linux necessarily.

> - RepRap host 0.9

good.

> - Darwin hardware (moulded or laser-cut?)

Are both of those kinds sale right now? In laser cut we've got Ian and Vik (shipping may be a problem. is it?) and, I gather, Zach (though I don't know when. Zach! when is your laser cut going on sale?)


> - some sort of extruder

which one is most reliable? do we know?

peteredworthy Wrote:
>
> Steve: I don't see why linux would be the prefered
> environment for the easy route or is it root ;-}

No particular reason other than it's a) open source and b) outside the Mac / Windows religious war.

peteredworthy Wrote:
-------------------------------------------------------
> I've thrown together a somewhat prescriptive page
>
> [objects.reprap.org]

Looks good. I've orphaned-out my build guide skeleton in favor of this one, to avoid confusion.

one could check wich OS is currently running on the clients machine, but that's a bit messy.

that being said, if we won't have the user to choose what he want's to build, because he might not know.
He surely knows wich operating system he has.
so why shouldn't he choose for himself.

that is: no recommendation for ANY OS but links to the big three.

I know I know.. Linux is .. and windows isn't.
But let's be honest the average person goes in a store, buys a pc and plugs it to it's wall... and what does he see?
So if it's one os to be "standard for the unexperienced" I think it has to be windows then. xp at the moment vista in a few month...
We don't like that? Sorry folks, then no OS but three links

'sid

I'm all good on these ideas including letting the user pick which OS to use from a list of three. That is a choice the masses are qualified to make ("which OS do I have?...this one...click!) but the software loading needs to be much streamlined. Can we do an installer and just bundle the stuff together in a nice big tarball and just worry about dependencies ourselves instead of making them do it?

I think the laser cut parts are the majority of the kits offered for sale so we should focus on those. For the most part I think everything can be bought in kit form from either Zach or Ian. That solves the sourcing issue and gives us a common set of components to build documentation from.

Anyone want to offer ideas or criticisms on the installer idea?

Demented

I am still against using the "builder's wiki".

I favor a brand new wiki which has a separate database and is completely clean when we start.

otherwise we will be scrambling to figure out what the heck we were just editing, which pages are with the new design and which pages are from the old design, what should link where and which version of a duplicate is the newer one.

please don't go this way it's going to be like a ball pit in there.

-Leav

degroof Wrote:
-------------------------------------------------------
> peteredworthy Wrote:
> --------------------------------------------------
> -----
> > I've thrown together a somewhat prescriptive
> page
> >
> > [objects.reprap.org]
>
> Looks good. I've orphaned-out my build guide
> skeleton in favor of this one, to avoid confusion.

Regarding your page. It looks good. A couple of observations, though.

I'd really discourage you from suggesting that people buy HDPE filament and stick with ABS. HDPE warps and shouldn't be recommended to anyone not prepared to work on the warpage problem. This has a knock-on effect. Which extruder is most reliable with ABS? Nop can help us with this, I expect.

Also,

"No pre-assembled option is currently available for the electronics and so you, or someone you know will need to solder the components to the board."

I talked with a hobby robotics company . They might well be interested in supplying assembled and tested Darwin boards and possibly even Darwin itself.

Sort of agree with Leav on this but I don't see as there is much else to do.

As to Assembled Darwin boards...That would be awesome if the price were reasonable. What are they talking?

Demented

Demented Chihuahua Wrote:
-------------------------------------------------------
>
> As to Assembled Darwin boards...That would be
> awesome if the price were reasonable. What are
> they talking?
>
They said they'd have to redesign the boards for surface mount, which they are able to manufacture more efficiently. The markup percentages for assembly and testing didn't seem hideous. This is all very, very preliminary and may come to nothing, mind.

sure sure. Surface mount would be okay, I guess. Since that seems to be a big intimidation point for new people coming on. Also, after a while electronics improvements will slow down and stabilize...

Demented

We're talking assembled and tested boards here.

Demented Chihuahua Wrote:
-------------------------------------------------------
> Sort of agree with Leav on this but I don't see as
> there is much else to do.

why? there is no technical obstacle.

I really think this should happen. and it's not that big a deal to do.
Otherwise it will all just be a big mess.

Edited 1 time(s). Last edit at 10/08/2008 03:40PM by Leav.

@Leav

Well...do you have access to the wiki and such to make the changes? Cause I don't. That's what I mean. Adrian says he wants it in the Wiki. Being the beneficent dictator for life (to steal a phrase from Python), that is about it for that.

Unless of course you want to host it somewhere else...don't know about that either. Don't want to split the project into documentation and other stuff.

Demented

no no i'll never do that.

I e-mailed Dr. Bowyer and we'll see where it goes, but I won't press the matter beyond that.


so..

are we focusing on kits? or is that just that they are easiest to add right away?

As Forrest Suggested, i'm building a Darwin Right now so i'll document as best as I can and take responsibility for polishing up the cartesian bot page.

Regarding format: I think I will break up the cartesian bot into several pages to make it more manageable. does that sound good?

-Leav

Sounds good to me. I can help out with polishing up the Arduino electronics pages as I have experience with that.

Hopefully it starts looking coherent and we can give people links to it when they get stuck on the other one.

Demented

Quote

We're talking assembled and tested boards here.

I think this is one of the best things we can do (at the moment) to get RepRaps into the Market. I think connecting the board together is less of a problem as many people (although not the mass market) are capable of doing this, or know people who can. Many people can assemble computers, and connecting the boards together is not too dissimilar to that.

Quote

are we focusing on kits?

I see two streams that we need to operate on.

1) The Kits

These are packaging the parts that make a RepRap in a way that someone with basic knowledge and skills can can use to make the RepRap. This is the "Screwdriver" package, where the only tool needed to assemble a RepRap is tools that would exist around any home (Screwdrivers, Hammers, Glue, etc), but if it could be achieved with just a screwdriver (or Alan key - or even snap together) would be optimal.

2) Instructions

This stream is about documenting the process needed to assemble the RepRap Kit (as in stream 1) so that the majority of the population would be able to follow it.

This would feed back into the Kit development stream as needs for simplifying the instructions could be achieved by changing the design of the Kit (and changes to the Kit will require changes to the instructions too).

Because both streams effect each other, both will need to be done simultaneously, or decisions made for one stream might have to be revised so that the goals of the other stream can be achieved. As in any engineering project (hardware of software), if you have to change a decision that has already been made, the more things that depend on that decision means that more thing will have to be changed and more work will be needed to make those changes.

Which reminds me of a joke:

How many software engineers does it take to change a light bulb?

None, it is a hardware problem.

How many Electricians does it take to change a light bulb?

None, we can fix it with software.

As for linux/windows/mac, I think there is only one "simple to install" answer. That's a bootable CD/DVD that comes with everything. There have been massive problems with java libraries and versions of libraries on ALL three platforms. It's a nightmare to get working correctly. Too many people have failed, and needed expert platform-specific advice to get unstuck.

It's not ideal, but I think recommending anything other than the bootable CD for a newbie is asking for trouble. The installation is just too complex, and there are to many places to go wrong, and too little feedback from the computer itself as to what needs done to fix it.

My prior experience with Knoppix leads me to think that one of the main modifications we might want to make to the CD/DVD to have a writeable area by default on the user's real hard disk. Someplace to save all your modified models, custom extruder settings, and so forth. Otherwise, knoppix has a nasty habit of forgetting everything between reboots unless you take great pains to tell it otherwise each time you reboot. I suppose that I haven't looked at knoppix in years, and that might have changed...

---

--
I'm building it with Baling Wire

I'm glad somebody else said it.

That sort of thing should be possible but I'm not sure how much of a change it would take to the default settings of everything. I'll have a look at it for the next iteration of the LiveCD.

I must admit,
I'm still searching for a viable path with the hooks mediawiki left for code additions.
A javascript is in almost working condition but it's not too small,
and leaves a bad layout if js is deactivated on the client machine

short: I'm not quite there yet

'sid

Sid,

I'm sorry if this sounds as if I am dismissive of your hard work, but...

are we even sure that the uber-complex branching is the behavior we want?

seems to make much more sense to me to have a modular layout and have one main page as the tree from which we branch out.

to fully utilize the tree analogy:
Trunk - There is/will be a main design which will be the trunk.
Branches - Alternate designs which some people might want to use.

The main instructions page would link to the tree, in a way which would make it obvious which of the designs is "trunk" and which is "branch".

Since building a RepRap is not a 10 minute deal, it would not be useful for end users (as far as i can tell) to have a "wizard" type tutorial since they will often want to back and forth.

this is something that can be solved by properly designing the "tree" page and making it easy to use.

-Leav

true, exactly that's why I think that's the best option here.

Let me explain why:

Take your tree, with lots of alternatives for every step in that tree:
day one: user takes the first lets say four steps on three pages
day two: user wants to go back two steps because he thinks he's done something wrong now he needs to remember wich option he chose or has to search around for the page that looks familiar.
day three: he thinks about a printout of everything he needs to print every step seperately searching around for the correct pages, reading their description and so forth.

Take my select first and then only back and forth:
day one: user chooses what he wants to do, for the best he can (all other will be preset default options) and he makes the first four steps on three pages
day two: as above he figured to go back two pages (just by hitting the "go back one page" link twice because of the optioncookie he has) and can be sure about having the correct page within seconds rather than minutes.
day three: he thinks again about printing all needed steps to have them handy at the shop where he doesn't have a computer, still he has to print every step
seperately but he only needs to clicks per page(next page & print page) because the website KNOWS WHAT he needs, no guessing no searching no description reading
(all necessary descriptions he read on day one and made his choice)

wikis are nonlinear websites, that's find for wiki-esque things like dictionarys, where one page stands fo itself and is not necessarily part of a bigger set of informations.
For building instructions where you simply have to follow ALL directives most likely in a certain order a linear webconstruction is the most convenient way for users.
Building a tree is fine, for us, for those who have an idea of what informations we(they) are looking for.
We can tell by "thermoplastextruder Mk2" that it is older than v1.1 (although it is named v2.0 in it's description) we know that it is directly driven (what seems to make more sense for some plastics) and we know which informations we need to grap there and which we don't.
But how should a Newbie know?
He can't he has to read everything, most pages twice take a close look in the forum to know what he should do best.
So why show him a tree structure with both extruders?
Show him the v1.1 and he's good to go (maybe a small hint like mentioned there for the improvements)

Most important (a least that's what I think) is: clean up the cartesian bot files at sourceforge, at least make another set called (beginners bot.good to go v1)
kick all the parts that are not documented (I'm still wondering about all the "caliper" parts) a set with NOTHING but the parts that are mentioned in the instructions everything else is just confusing people.

Don't get me wrong, I don't insist on coding a plugin or writing (cleaning up my) javascript, lots of other things that can be done around here.
But as far as I know internetusers are not only patient or even reading people.
think about all the semistupid posts in the forums asking things that do not belong here... *breakingbonesound* for software we mentioned, asking about how to do xyz with a microcontroller somebody mentioned and so on;
People are reading few words that match and then they make the decision to post.
Same in making decisions on "what do I do"
If they see a tree with lets say twenty entries, all they think is:
wow, that's along way to go.. many pages to read... I'm outa here.
if it's one page with few options to make and they're left with seven or less pages, they'll give it a try.
hence why almost noone tries to build their own "cintiq" (sorry for misusing a brand) too much steps to take, although it's easier to do than to build a darwin.

We must be small but effective (on informations) to those who just want to fly in and try out.
No need to abandon the full set of informations for all those (like you and me) that are willing to read many days to know more.

whoo long post above there.. I'm sorry
-quit-

'sid

well that stretched my attention span to the limit

I agree on the need for a simple, single, set of instructions, and I agree it should be adapted to the "internet attention span".

I still view this as very possible to do within the "confines" of a wiki.
this is simple, and is already used in wikipedia in several forms:

-------------------------------
You are here:
.....V
[1][2][3][4][5][6][7]
-------------------------------

is fine, together with a zeroth page, which explains the building process and includes links to other possible pages:

-------------------------------------------
Blah blah you are about to build a darwin blah blah blah.
bookmark this page so you can always come back here.
here are the major steps using the standard design:

1 get parts
2 ???
3 Profit!
4
5
6
7

Variations:
here are the basic variations that exist, with a short explanation. click on them to learn more. they might save you work if you are in a certain condition.

a one line
b each
c so it
d remains
e readable
------------------------------------------



once they click on [1] they are sucked in and can not deviate from the line (this is good!)





It is inevitable that they will have to do some reading around, because Darwin will never be an "HP Deskjet 650c" that just plugs in and is intended for people who triple-click icons.

Later RepRap versions will perhaps set that as a goal, but currently I do not think it is possible, or desireable.

this is going to be a short one...

Quote

It is inevitable that they will have to do some reading around, because Darwin will never be an "HP Deskjet 650c" that just plugs in and is intended for people who triple-click icons.

True!
But to spread fast and efficient we need to get as close as possible
Maybe not to HP Deskjets, but maybe to LEGO buildplans.
at least a numbered order of where and when things are build.

'sid

yeah I think LEGO instructions are a great example!

(Edit: for what we should strive for in the simplicity of our build instructions)

Edited 1 time(s). Last edit at 10/19/2008 01:25PM by Leav.

LEGO build plans...
[objects.reprap.org]

Sweet! you actually used MLCAD

I think that this could be a great idea.
it doesn't have to be CAD (though that would be nice in my opinion), but a the general style of how Lego conveys it's instructions in such a graphical way as to not even need words is a great thing to strive for.

the even better thing is that if we strive for no words, and then use words,it shoud probably be the most coherent manual ever written

-Leav

Quote

Maybe not to HP Deskjets, but maybe to LEGO buildplans.

I agree, RepRap is inherently complex and it would take a lot of development to get it to the ease of use of a typical desktop printer (not that we shouldn't try for this).

As a first step to this, I was thinking along the lines of a flat pack furniture product. Basically, all the components are included in a package and all they need are simple household tools like a Screwdriver or a Hammer (well a hammer wouldn't work for a RepRap ).