Encouraging development 2 - Incentives

I've noticed quite a bit recently that reprap developers (and I am as guilty as everyone else) have a ridiculous amount of hubris. The "I can do better" attitude, cousin to "how hard could it be", which started the platform to begin with. Attempting to presume a standardized distance for all x carriages, by way of example, is an extreme exercise in futility and imposing the hubris ever more.

I think nophead's approach is the most user friendly in design strategies allowing those streaks of "I can do better" to be massaged if desired. However, Bubacks argument seems to suggest nophead plans for the mendel 90 to be the one design to rule them all which I don't see.

My point earlier is that unless handled with kid gloves by a very small and very dedicated (even insightful and lucky) locked-in core team then standards and impositions placed on an open source project will stifle creativity and limit the chance for a projects greater success. Consider the evolutionary theme given to this project by Dr. Bowyer - time will be the true judge of what design elements win out in the end. If a design is truly successful other designs will start to fade away. This very thing happened to the Arduino in that all the fakes while still out there are not sharing in the same success that the primogenitor is.

Maybe I have misunderstood something but it seems that rather than create a standards base development process, we all benefit by encouraging diversity and growth and contributing not through top down design decisions but instead by making - making improvements, making derivatives, making more printers, and making new designs.Yes that might mean 50 different x carriages for the time being but give it another 4 or 5 years and this will inevitably begin to narrow.

It's not as complicated as all this, I don't think anyone is trying to define a standard at this stage, this discussion should hopefully be leading to feedback (survey / comments) on all the masses of existing designs and printers out there. Their weaknesses and strengths, highlight mods and changes people have made (both released and work in progress) - and maybe if we are lucky a few interesting ideas and wish-lists people would like to see in the future -

Then when we have all this maybe start with some of the weaknesses and consider the alternatives.

*- For example the 8mm smooth rods issue, maybe all 'new' designs should be able to handle both 8mm or 10mm with a simple insert or reversible clamp? That's not defining a standard, just saying that 8mm rods can bend on long carriages, 10mm works better, but can cost more so use the 8mm rods you already have and upgrade to 10mm by just reversing the clamp if your printer grows in the future.

Then maybe some of the more interesting modifications -
Dual Extruders -
We need, a dual carriage, Firmware, slicer support, modeller output (maybe STL is going to limit us doing this well) - all good stuff we can then discuss.

Then finally we could consider the more contentious issues, like choice of frame style and material - and most probably we will still decide that you can build it however you see fit, in whatever material you like.

We are trying to build an open knowledge base, not a closed one, the one thing we don't seem to do enough of is share our discoveries when we discover them, it's usually only when someone else posts that they are having a similar problem.

This process does not need to be hard, it should be the exact opposite and we should all learn more from it, both to help ourselves and the 3D printing community. Can we at least try to gather pro's and con's in a honest and pro-active critical way.

---

[richrap.blogspot.com]

Perhaps the word "standard" is too hard. How about "recommendations"?

After all, it doesn't make sense to have one pair of rods 50 mm apart while the next one has 55 mm. Still, with a recommendation there's nothing stopping one from making them 100 mm apart if this gives a technical advantage.

We already have such recommendations, just not written down. For example, to use bipolar stepper motors. Unipolar ones would be quite doable, but electronics would no longer fit. The usage of G-code is a standard. Unfortunately all those extra G- and M-codes are quite a bit messy, as they aren't discussed. Discussing these extensions would help a bit cleaning up the situation without doing any harm.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

Exactly. People can, and do, build non standard machines.
Having a standard (or baseline, scientific control, benchmark, normalized, common, stock, whatever word you want to use) places no restrictions on those individuals that want to innovate.

(I apologize if I'm harping on this, as it's a digression, but I would like to see a baseline established. if somebody doesn't agree with the choices we decide upon, then they don't have to follow them. It's not like we're making laws here.)

anyway, I'd like to see a graph of print area practicality. something like cm^3 over (plastic cost for the total filled area), or cm^3 over print time. something like that would help us identify the bounds for a practical machine. It would also help us compare various alternative build methods, like powder bed and binder fluid, or powder bed and sintering laser or UV epoxy and laser, etc.

Correspondingly, i think that is the real power standards; we need some baseline machines in order to do some quantitative comparisons.

the survey is another piece of the puzzle. If people really want a large machine (and some clearly do), then we need a way to make it practical. Maybe "the masses" want a small bookshelf machine, or a cheap-and-simple machine. Standards and surveys are just data points for those interested in them. I, for one, am eager to see the results

The whole point of having, for example, known "standard" distances between axes is to be able to exchange key parts of design A with design B, should you find that A is not up to the task. "Interoperativity".

As a user, I do not really care about who or what process generates that, but I do benefit from having a variety of choices without having to reprint or refurbish the whole gizmo.
On the other hand, the more "unique and special" a design is, the more Inclined I'd be to do the mods myself, and to hell with publishing the mods: limited exchange and diffusion, the medieval way.
Some degree of interoperability would help create the market.

Manufacturers and marketers have fought "bureaucratic, imposed standards" since times immemorial yet much of it is orented to generating a real market for consumers, and this market expansion has in turn obviously benefited manufacturers and marketers.

I have now some doubts about the usefulness of a survey if makers would just see it as "bureaucratic" stuff. Either you believe this stuff is automagically self-regulating (many industries advance similar pretensions), or you think some degree of planning and concertation might help overcome certain limitations. But then no one would want to pay for that overhead.

My take is that as long as there is massive information assymetry, this won't converge to equilibrium without some organization, culling, and, to say it bluntly, policing. The scene reminds me a lot of the early nineties when clever boys would overclock 386s and 486s and re-sell the stuff (undeclared as overclocked) to users-suckers, because "to run a word processor you don't really care". More than one made his first million that way. Of course machine crashes and unstabilities were blamed on "that crappy OS". In some cases the processors were summarily tested with loop-running Doom sequences. Yeah, entrepreneurship at a peak of creativity.
When I see some hardware being sold in the reprap ecology (fortunately not the majority fr now AFAIK!), I have that same old impression.

Having done some of the work to get a vertical X axis standard put together, I agree with what the last few posts have pointed out, and I might even go a bit further. There may be some worry that creating standards will suppress creativity, which I can't imagine considering the new designs coming out every day or so, but there is another side to the story. The reprap community is now supporting a fairly large marketplace as well. It's not just design space any more.

Lanthan did a good job describing the effects of standards on the marketplace, but I don't think he went quite far enough. The truth is standards benefit only the consumer. Designers and retailers benefit from specialization and product lock in. Yes a standard expands the market, but it also brings prices down and causes increased competition and comoditization (Which retailers hate). So if we really care about bringing 3D printers to the masses then standards are a good step in that direction. If on the other hand we care more about making money for the designers and retailers then it's better to have a bunch of incompatible competing machines.

On the other hand we're never going to see any very restrictive standards anyway here in RepRap land, so I don't think talk of standards is a threat to anyone. Everyone is free to follow a standard or not. The only pressure would be from consumers if a standard is adopted by a large number of customers. And that's the good kind of pressure, it means lots of people want reprap stuff.

And since this thread is about incentivizing development, what better incentive can there be than broad compatibility of a new design with the existing customer base. If for instance there were a standard describing basic compatibility rules for hot end mounting, then a designer could design for the whole reprap community rather than having to produce 2 or 3 different designs for the different extruder types out there. That is a big incentive right there.

One thing that I am struggling with, as a newcomer, is trying to work out which of the experiments out there are worth incorporating into my own machine, and which are invalidated or failed experiments.

For example - what am I to make of this? [www.thingiverse.com]

It would be great to see more of the incremental improvements that various people have come up, and the community consensus as to their value, make their way to the wiki. Is this a desire that others share?

> Having done some of the work to get a vertical X
> axis standard put together, I agree with what the
> last few posts have pointed out, and I might even
> go a bit further. There may be some worry that
> creating standards will suppress creativity, which
> I can't imagine considering the new designs coming
> out every day or so, but there is another side to
> the story. The reprap community is now supporting
> a fairly large marketplace as well. It's not just
> design space any more.

One way to deal with the chicken and egg problem of getting a standard established is not to call it "the standard", but merely "a standard". For example, bryanandaimee might come up with a set standards that define "B&A Compatible" customisations. The challenge is then to get that standard adopted. No-one is going to be forced to use your standard, and so you need to make it easier for people to adopt. Things that will help:

* Make the standard match as many of the most promising customisations as possible
* Catalog the customisations that already meet it
* Talk to the people who made those customisations to let them know that you like some aspect of their design so much that you want to make it a standard
* Create variations of incompatible yet promising customisations so that they meet the standard. Contact the creators of the originals, explaining what you're doing, and seeing if they buy in.

Someone who is interesting in developing standards should just make one - well, a v0.1 of one. Discussing it in the abstract is unlikely to go very far. Having a specific recommendation that people can say "I like it" or "I don't" will likely do much more good - best case scenario, you can take on the recommendations that people come up with before making v1.0.

Quote

One thing that I am struggling with, as a newcomer, is trying to work out which of the experiments out there are worth incorporating into my own machine, and which are invalidated or failed experiments.

Some things work for some people and not for others. I experimented with lever based extruders long before Greg published his. My levers all broke so I abandoned the idea. They obviously work for some people as his design has become very popular.

---

[www.hydraraptor.blogspot.com]

Interestingly, the main comment I get from designers that are using the vertical X axis standard is that the standard is not detailed enough to ensure compatibility of the various designs. We didn't specify things like belt anchor points, and X to Z bar spacing. That was intentional since we wanted to leave as much freedom as possible while still hoping for basic compatibility, but I thought it interesting to note that the designers are often the ones that want standards they can design to, and push for detail in the standard.

Here's another thought, for those who spend their entire life on reprap and create full machines from scratch like mendel90 and mendelmax, a standard is not as usefull, but for those who might spend a small amount of time hacking on an extruder or something, standards really are a benefit.

Standards are always useful.

Since I've iterated a few designs over the months a surprising number of minor (though important) design decisions are purely arbitrary. Having some standard simplifies those decisions, and also makes things more compatible.

---

www.Fablicator.com

Someone put this in the comments section of that Hack a day article on saving reprap. Seems to also define the technical politics of those in favor and against the notion of developing standards.

Its interesting that this discussion is being posted in a 'Develoment' section.
I think it belongs in a 'Standards' section.


To quote the always useful WikiWikiWeb:

“Standards and methodologies don’t work!” says the developer.

“How would you know?” says the Standards and Methodology Guy. “You have never followed them.”

“Neither have you.” responded the developer.

Quote
buback
Aren't you implicitly defining a standard? It is based on standard parts and it is all a defined algorithm in your OpenScad file. Others will need to refer to your .scad file in order to make alterations. (and it basically restricts designers to using openscad)

That's EXACTLY how the Prusa Mendel (the one you recommend as a standard, remember?) works. You probably ought to know that.

First of all, This is silly "This entire page is a horrible, terrible idea, and you should be shot for creating it."

I have no ill will against anybody in this forum. If you or anyone thought my words were heated or too antagonistic, I apologize.

Regarding your point: I don't mind if it's an openSCAD standard or a blueprint plan drawing. My point was that even a .scad file describing a parametric machine is still a standard that others have to base development off of. I don't see anything wrong with acknowledging that fact.

I would prefer some firm numbers for certain variables. However, that doesn't mean I'd suggest a newbie build Sells Mendel, and I have no problem suggesting Mendel90, even as a "standard" reprap (once there are a bunch out in the wild).

Buback Wrote:
-------------------------------------------------------
> First of all, This is silly "This entire page is a
> horrible, terrible idea, and you should be shot
> for creating it."

I saw that too - It's beyond silly - it's infantile and unwelcome. I just deleted it. I had issues with calling it a Standard, but harassment or name calling is a truly stupid way to participate. Stable Build is fine w/me.

Regarding incentives and standards - I tend to agree that as a new developer, I'd be more inclined to read standards (so far as they pertain to cross-compatibility) and participate in their development on the wiki. A proposal is much more effective than a request.

That said - I consider myself a fairly experienced engineer - I'm going to make design decision that very well could rule out conformance to all standards anyway. But - if I don't have some well documented reference to consider going into the design process, you can be sure I'm not going to download & build every other design out there in an effort to conform - honestly it's just too much work for too little reward.

sminnee Wrote:
-------------------------------------------------------
> One way to deal with the chicken and egg problem
> of getting a standard established is not to call
> it "the standard", but merely "a standard". For
> example, bryanandaimee might come up with a set
> standards that define "B&A Compatible"
> customisations. The challenge is then to get that
> standard adopted. No-one is going to be forced to
> use your standard, and so you need to make it
> easier for people to adopt. Things that will
> help:

I like this approach a lot. It puts the onus on someone to submit a proposal and it doesn't claim to be king - it just points out what's required to mate up to a design (or set of designs). I'd add that in addition - if you want to submit a proposal - you might add in some design methodology too so that "intent" is understood as well as implementation. For example, "On the x axis I located the motor opposite to the belt with the gears between to maintain as close a possible a neutral weight distribution and avoid unwanted torque loads on the bearings." This could be far more useful than simply stating where something is in maintaining future compatibility.