Re: Betr.: Re: Rename strange/unexpected behavior.

Hi Alexander,
 
Thanks again for your comments - they are quite insightful.
 
I think you have got to the heart of the issue that when the ratio of consumers to contributors gets too small that it causes problems.  There doesn't seem to be any chance that we'll run into a problem of 'too few consumers' any time in the near future - CVSNT delivers SCCM features that no comparitively priced (or zero-dollar) solution offers so I already have more sales calls, quotes and enquiries to handle than I've got time to process .
 
However I believe that by introducing a paywall we are setting a very very low bar for consumers to become contributors - each consumer hands over £85 and no need to do any more to ensure that the community has enough resources to continue to develop and support the products.  Some people automatically see this barrier to entry as too high, eg: if they see the cost as higher than any perceived benefit.  
 
This also gains us an enormous benefit of excluding people who were never going to invest enough time/energy in implementation to recoup those costs and for whom an alternative solution would be a better fit for their requirements, ie: excluding people whose only reason for choosing CVSNT was that it was zero-dollar.  
 
IMHO: Price should never be the primary factor in choosing an SCCM toolset - you need a business driver, a SCCM methodology that addresses the business requirement, and a toolset that supports that SCCM methodology.  At it's core SCCM is always a 'cost' (or overhead) so you never really get zero-dollar SCCM, and if the problem that SCCM is solving for you, costs you less than £85 then CVSNT is probably not the tool to solve it with.  IBM have the same attitude with ClearCase (which I think is a terrific SCCM tool - albeit a very complex and costly to administer terrific tool), they price it at around £5000 per person because if the problem that SCCM is solving for you costs you less than that then ClearCase is probably not the tool to solve it with (partially because it introduces high overheads).
 
BTW - I personally feel that you describing Hg and Git as 'flashier' solutions is amusing.  Based on my above definition of how to choose SCCM flashy isn't a factor.  I and the CVSNT project have always been primarily focussed on tools that save time/money and solve specific business problems (something that can be described in non-technical terms, something meaningful to the company who employs you).  
 
For example one customer (software vendor) a few years ago chose our solutions because the business was losing customers to an alternative vendor who could supply patches for individual customer bugs, wheras they could only supply 'releases' (service packs of mulitiple customer fixes).  So there was a clear measurable cost of not having an SCCM process that supported fixing bugs in isolation (so if two bugs affected a single deliverable then only one bug at a time could be worked on).  The methodology then required a tool that could manage multiple checkins across multiple modules/files with a single 'user-defined' peice of meta-data (the bug id) then create builds based on a 'baseline' plus 'user-defined changeset'.  CVSNT has this feature, SVN, Hg, Git etc do not.  The programming staff were adamant that they wanted a 'flashy' 'new' tool like SVN, Hg or Git - but what they got was CVSNT, because it solved the problem and the other tools did not.  I think that in the end the staff were happy enough about using CVSNT, it was flashy enough for them ;)
 
We don't have this basis for measuring the success of our products for our consumers for no reason.  And indeed Git and Hg and SVN have their own measures - each of these tools do not exist and are not as popular as they are simply because they are/were fashionable - they each solve a specific problem - and if that problem is costing you money then it's worth the investment required to get them running.  CVSNT is not the fastest SCCM system available (not that it's slow), but for one of those other tools speed is the ONLY thing important, if any feature slows it down then it's not added.  I think in 11 years of working primarily on SCCM I've maybe seen 1 company for whom such a requirement would have any meaning, and certainly not for me.
 
A long time to get to the point (sorry) - it's perfectly OK if some (many or even most) people don't choose CVSNT.  I want people to choose CVSNT not because it's zero dollarrs, not because it is or is not flashy, but because it solves a problem that was costing them money.  The majority of people who will be 'turned away' from the CVSNT community by the paywall will be people for whom CVSNT was not the ideal (or even a good) solution anyway.  Yes, for a (long) time there was value to be had by having them as pure-consumers - better to use 1% of CVSNT and they can 'grow' into it and have zero-dollar cost of transition as they grow, but unfortunately allowing these people in caused other people to be prevented from contributing and the cost of those lost contributions was too great.
 
There clearly is a self sustaining community outside of March Hare for CVSNT, hence why we are discussing it here, and even why TortoiseCVS exists at all.  The good faith you've shown towards others in the community is noble - but I suspect that for quite a while I'll be seeing accusations, including "the CVSNT community is dead", from people who want to still use CVSNT not some other system and who are upset they can't get the software for zero-dollars anymore.  I feel for these people, they've invested CVSNT into their own or employers business and are fearful that maybe someone will now criticise their decision - or it's going to cost them money and/or time that they never expected which may need to come out of their own pocket.  I did not take this decision casually, and took it under advice from FSF and those most active in the community and only after trying ever other possible alternative.
 
We are in a transition period where people are still getting used to the idea that to run a CVSNT server you now need to pay a fee, it's a recent change and it'll probably be another 6 months before it's settled down.  I've responded to questions about it here (even though strictly it's OT) for many reasons, but probably primary two are because I'm interested in feedback and dispelling FUD.  
 
Finally: one of the problems of 'mailing list' communities is they tend to be very quiet when the software works well, is easy to use and is documented well - because there are no bugs and no questions and no misunderstandings.  In the next 6 months at March Hare we'll be trying to build much more discussion with our customers about SCCM process, hint's and tip's, training etc etc. 
 
Thanks again for your time and valuable feedback.
 
Regards,
 
 
 
Arthur Barrett
 
 
 
 
 
 
  

	-----Original Message-----
	From: [email protected] [mailto:[email protected]] On Behalf Of Alexander Yuan
	Sent: Wednesday, 26 January 2011 6:19 PM
	To: Arthur Barrett
	Cc: Jan Keirse; Dries Feys; tortoisecvs-users-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
	Subject: Re: [TortoiseCVS] Betr.: Re: Rename strange/unexpected behavior.
	
	

	On Tue, Jan 25, 2011 at 10:01 PM, Arthur Barrett <arthur.barrett-qn/[email protected]> wrote:
	

		Thanks for writing - sorry I didn't catch your name?


	Sorry - my name's Alexander Yuan.


		
		
		
		I paid the wages/costs of CVSNT development staff for over 6 years at a cost to me personally of at least £350,000 (probably double that), and I do not own a home or other investments, I have no benevolent provider for my home or office rent electricity and phone, and my car is a 7 years old bottom of the range Volvo.   I gave copies of this software that we developed away under no obligation and it led to specific problems that the FSF advise a paywall solves.  What you are suggesting has already been tried and failed, and now I'm trying a new way.  This new way is already successful but maybe in another 6 years we'll need to replace it with something else.  Right now - all around me I see F/OSS projects introducing 'paywalls' and all people need to realise - if you don't pay for free software (as is the FSF's advice) then in the future you will not get free software, you will only get proprietary software.  There are real tangible benefits to using Free Software and it has nothing to do with price.
		  
		There was a time when not charging for CVSNT was beneficial - but that time is long passed, the article I published at the time goes into much more detail - see the sections "Why is SCCM with zero-dollar price tag a good thing?" and "Why is SCCM with zero-dollar price tag a good thing?":
		http://www.evscm.org//modules/Downloads/
		 
		CVSNT will be going for many many years in the future - continuing to lead the way in Free/Open Source Software Change and Configuration Management for commercial software development teams as we've always done.  Do you want to join the project today, or in 5 years time after wasting 5 years on tools that don't deliver tangible measurable benefits to commercial software devlopment teams?

		 

	
	
	I think there's been a miscommunication here.  My surprise was at the claim that the community grew more active after Mach Hare instituted the pay wall - but I don't see how this is sustainable, at least with the type of community I am familiar with.  Let me make things a bit more rigorous with what my mental model is.
	
	
	I noted in your first e-mail that you disliked that users were requesting resources from the community, but not contributing back through development or support help or the such. In your FAQ, you say you feel that distributing CVSNT gratis was harmful to the community because there were too many people whose involvement were net losses to the community.  So we have users who are consumers, users of the product that only consume resources from a community.  And we have contributors, those in the community who are the volunteers, donaters, enthusiasts that provide a net positive amount of resources into the community.  In this model, the problem is that there are too many consumers competing for too few community resources.  If I understand correctly, you've decided to address the problem by raising the bar enough such that only those who already are willing to be highly invested will get community resources.  This essentially limits the consumer pool to only those who are already willing to be contributors.
	
	
	The FAQ mentions one positive effect of distributing your product gratis - "reciprocity, generosity, and philanthropy".  My point is that communities that revolve around free community-contributed code (or content, or support) sustain themselves through precisely this effect.  The example of Wikipedia was to illustrate how a community can be successfully built on that pillar.  Some consumers eventually become contributors, who improve the quality of the free product and bring in more consumers.  The community can grow.  If the trajectory is right - that is, if enough contribution of content occurs that eventually there are more contributors than there were - it becomes a virtuous self-sustaining cycle.

	Somehow, in this case, you seem to have encountered a terrible ratio of contributors to consumers, or low contribution per contributor, or poor signal-to-noise, or something of that nature.  But then the solution is not to institute a pay wall, which drastically reduces the number of consumers - they are the source of contributors, and removing them makes a sustainable community impossible.  This seems particularly true in the case of CVS, which already isn't getting as many new consumers compared to flashier solutions like Hg and Git.

	In essence: in this model of community-building, a pay wall can lead to too few consumers, and the community will shrink for lack of contribution.  I see how the barrier to entry can improve the quality of the community, but not any way it can sustain its size.  Personally, I cannot think of any other model of building open source software communities that is sustainable.
	
	
	I do notice that there are financial reasons for why you are charging for distribution, but one would not expect the typical self-sustaining community of volunteers and enthusiasts assisting you.

	That should hopefully explain why people might not consider there to be much of a CVSNT community.  It might also explain why people expect there to be some sort of self-sustaining community existing outside of Mach Hare.


	

		The only people complaining about the changes to the CVSNT project are those who are not willing to contribute which I'm happy to ignore, and Wikipedia likewise pays no attention to anyone unless they contribute.  It's all about contributing.  Wikipedia define contributing in a way that makes sense for them, and the free software foundation define contributing in a way that makes sense for Free Software, and March Hare define contributing in a way that makes sense for CVSNT.



	To be precise here, I should note that Wikipedia contributors are (by rule) supposed to write to the benefit of all readers and not just other contributors - for example, English articles are generally written for an international audience, even though almost all editors are from English-speaking developed countries. But that is a bit of a tangent; my point was that Wikipedia successfully offers itself for free to build an audience, and uses its audience to build its community.

	( Actually, in the case of Wikipedia, they may be overheating to the point where they are burning through their contributors faster than there are new Internet users from which they'd eventually get more contributors...but that's a story for another place and time. =P )

	Alex

------------------------------------------------------------------------------
Special Offer-- Download ArcSight Logger for FREE (a $49 USD value)!
Finally, a world-class log management solution at an even better price-free!
Download using promo code Free_Logger_4_Dev2Dev. Offer expires 
February 28th, so secure your free ArcSight Logger TODAY! 
http://p.sf.net/sfu/arcsight-sfd2d

_______________________________________________
tortoisecvs-users mailing list
tortoisecvs-users-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
https://lists.sourceforge.net/lists/listinfo/tortoisecvs-users