[Eg-oversight-board] Release manager, Package releases vs Software releases, Membership fees

[rogan at roganhamby.com]

> Aug 15, 2013 11:52:18 AM, stephen.elfstrand at mnsu.edu wrote:

> Hi Kathy,

> Thanks for your response to my suggestions.

> I didn’t generate a lot of discussion with my suggestions. Yours is the only response I’ve received. I guess I’m just out of step with the thinking of the rest
> of the EOB and the developer community. 

 I responded earlier but it may have been lost. I'll try to repeat my thoughts here.

> I hope everyone takes my comments as constructive criticism. I do appreciate all the hard work and professionalism that the developers and the community at
> large have shown. I realize that many great things are going on. My goal is to take it to the next level.

> As an administrator, I would like to see
> Package releases not just software releases. For example

To be clear, and this is semantics but important ones, Evergreen does package it's releases. Non-packaged releases are a very different beast. The issues you bring up below are QA issues. I think the board could have a valuable role related to QA. I think one of the first is to define goals. You can't say "perfect QA under all scenarios." You might be able to say "Installs with no errors on a supported Ubuntu release with suggested dependencies." Galen will probably cringe at the overly simplistic nature of that but ...

> _At PALS we run only standard releases, and have had upgrade scripts them fail.  We try to exercise due diligence and figure things out for ourselves before
> we burden the Community with questions but have had  to go to the developer list and see if someone might know why it would fail.  We’ve had to find someone who knew why and apply an 
> undocumented fixes. This makes it difficult to treat Evergreen upgrades in
> a routine manner and adds unnecessary costs to us and service delay to our users.  It makes one wonder whether the upgrade scripts are adequately tested. 

> I’ve suggested before that someone not involved in a recent release should run the upgrade scripts to make sure the instructions are complete and correct and
> that the scripts work. PALS would be willing to be one of these testers..

 I'm not aware of anything that would stop any of your staff from doing this now. In fact if more did and gave more feedback we would probably have fewer such issues.

> -Acquisitions was released before it was usable, it was much too slow for large invoices.

True. I would like to see this a different thing but can the board actually hire and run a full testing division? Some responsibility does come down to the users to test. SCLENDS was one of the entities disappointed by the original acquisitions. We were also part of the group that funded the new work. There is still more to do, some is even probably critical to some users though not all, but we've made large leaps forward no matter how you cut it. And it is usable now. Where were the users who decided on that development design? Did they test? Did they give feedback? Where were the other users even afterward? It took a long time to gather feedback to have a real image of it's issues. A large testing division can avoid these but it takes a lot of resources. A lone tester (often a systems person) can't duplicate something as complicated as acquisitions and test it well. That takes a lot of investment. 

> -Authorities was not working well yet was released. Batch downloads of new authority records should change all the records linked to it. I know for instance
> that we lost a highly prestigious customer because, at least in part, the authority control system was not working.  They went to OCLC WSM instead.

>_Technical and end user documentation is way behind the code, yes it is getting better thanks to the community. Still, being forced  to consult  2 or 3 versions
> of incomplete documentation is just not good. If we held software releases until the documentation was completed, it would give the community motivation to get the documentation  done.

 Yamil spoke to some of these issues today. I'm in favor of the board trying to help find a solution but no, I'm not in favor of holding software for the documentation to be ready nor in the board deciding how the software is managed. As I've said before nor am I in favor of us choosing the release manager. There is a difference in F/LOSS culture and more centralized development. Part of this is separate community responsibilities and meritocracies in action. I no more think we can dictate to the developers than the developers can dictate to the web team. Anyone who wants a say in each group can join and do work in each group.

> Giving the Board some money to use for critical development would be helpful I think . The Board could supplement what is being done voluntarily by the community
> and fill holes by sponsoring development and documentation. 

This will be tricky but I agree.
 
> “;only one person submitted his name as a candidate for the position”
> Yes, I imagine it’s a lot of work to volunteer for. If the Board had some funds, it could offer a stipend to the release managers and get more candidates.

I would support this as I would someone to help backfill documentation.

> We talked about the need for funds to for instance, to use to match grants we might get. One idea is to sell advertising space on the Evergreen web site to vendors. Another idea is to have a 
> membership fee.
> Kathy you’re right to say that we shouldn’t charge people to run Evergreen. I’m not suggesting that.

I don't like the advertising per se but I do like the idea of sponsorships. We could call them memberships.
 
> As a public entity I can pay a membership fee but I just can’t donate money to worthy causes. So if there was a membership fee, I would pay it.  I guess to
> flesh out my idea a little I think a membership fee should be tied to voting. Organizations and individuals who wish to vote on community business should be members, so should the EOB  members. 
> Remember we pay 10% of our total billings to the Software Freedom
> Conservancy, I don’t see a problem recouping that money back through membership fees. I’m not talking about large amounts but something would help.

Pretty much every public library will have those same limitations, myself included.
 
> Crediting organizations for in-kind efforts does seem fair, but unfortunately doesn’t help raise actual dollars to match with grants or sponsor development
> and documentation.

Or grants, which will need some money on the table, for many of them.

> The fact that Equinox is requiring documentation is its contracts is great. I think every contract should include it. Let’s set this as a policy or best practice
> suggestion  as a board. Let’s have a documentation hack-a-thon.

The doc folks are already working on their own event which I think is great. I would be against trying to create any sort of policy about development contracts (and it would be impossible to support) but a page on the site listing best practices and maybe high lighting those vendors under the newly adopted policy support those best practices may be a good idea.

>We are competing with the Next Gen LMSs as well as current ones. The next gen systems are including knowledge bases for serials, OpenURL linking, and ERM  in
> an integrated work flow,  as part of library management. We need to move in that direction.  

I know many would disagree with you heartily on this one but I agree with you. I would love to see a stacked release of Evergreen with produts like coral and d space and heck, wordpress as well as apache/postgres.

> _I think Evergreen needs to be able work with the OLE folks to use Open Knowledge Base and incorporate these next gen features.

 Not critical to me but would be nice. Polling the community is going to be critical for the board backing new features. I'm more inclined to use the board's resources to support grants to bring in development on new features since it's expanding the community's resources. If we only use the community's existing resources I think the bar is raised for how much of the community will feel they benefit from it.

> _Academic libraries

> _ need a patron API to keep our patron database in sync with campus registration systems before large will want to use it en mass.

> __effective Authority Control

> _ interaction with Campus financial systems

> _NCIP interactivity is critical to all of our potential customers in Minnesota, without it they just won’t consider Evergreen. PALS has undertaken to help develop
this.

> _ Large public library systems will not move to Evergreen without

> __a first class Acq system

> __effective Authority Control

> On the Discovery side we need to incorporate

> _fast SOLR based retrieval 

> _come up with a framework for OPAC layout that will handle mobile devices as a matter of course.

This is an ongoing active discussion about resources. If I had the seed money I have grants we could go after right now to help accomplish this and bring in the skills we need.

> _The OPAC should have the  ability to use APIs from 3rd parties like Summon for Web-scale discovery.

> VuFind does all this already. Why not package VuFind with Evergreen? Do we really need to create our own OPAC?

I can tell you that for a number of reasons I have no interest in VuFind. It's a fine product but I could not make it a priority.
 
> Or if we do, why not use a CSS framework like Bootstrap, which handles mobile devices and is easier to customize.

> These is not just a theoretical concern for the future, We’ve already lost several prospects to OCLC Web Scale Management.

> PALS is willing to work on development, quality issues and strategic priorities, It just frustrates me when we have to spend time figuring out things like upgrades
> that should be routine, or compiling documentation from multiple sources, or features that are released before they are ready for prime time.

> So the reason I want the EOB to have funds available to it, is for it to sponsor
> strategic development that may not be of immediate concern to existing users but will position Evergreen to compete with the Next Gen Systems being developed. This will help appeal to potential 
> customers not just current ones.  Appointing the release
> manager and paying a stipend will give the EOB more influence on development directions and ensure that releases are packaged more completely. Documentation and bug fix bounties could be 
> sponsored as well.

Strategic development is exactly what I've had in mind when discussing grants and fund raising. I worry it won't be enough which is why I support the sponsorships.
 
> Maybe what we need is
> product management in addition to a release management. If the EOB had some funds and more influence, it could play that role. We need to set long-term development priorities to make Evergreen a > force in the market of the future and go beyond its’ reputation
> as a very basic system for small libraries.

I think the board can take an advocacy role and try to help with the long term path but the phrase "product management" concerns me. I'm not interested in running an ILS company. I'm very interested in being part of an open source ILS project.

Plus, at least here in the south east Evergreen is not known for basic functionality or being for small libraries. We have a different reputation here at least.

> We’ve talked about using some crowd-sourcing software to gather support for specific development and that is a good idea. The EOB could throw some money into
> the crowd-sourced pot for specific projects, if it had some.

I agree, with the caveats I gave above.

> It seems that the common opinion is that we really don’t need to change much because things are going well as they are. While a lot of great things are happening,
> I think we could do better with a little more long-term strategic management of the Evergreen product. That seems like the natural role of the EOB. Of course the EOB would always  consult with the 
> whole Evergreen community as it does this work.

I'm not sure where you have that opinion from. I work with a lot of the community and I don't have that impression from folks. The sense I have is very different. The sense I have is that a lot of folks don't know how to make change and there is some inertia from a period of complacency. I think we're getting past that now and I want to continue that. I don't have all the answers but I like to think I'm at least helping towards a better place in what I do. What I do is sometimes pretty measured though because while I don't want to lose momentum going forward not do I want to roll over what is already good in the community (much I see improving) and what is good about a F/LOSS community.

Stephen Elfstrand