The Gordian Knot

Enough serious stuff for this blog! I visited David Pattern’s blog the other night and idly gave the 2.0 Idea Generator a workout for some inspiration. Coming the week after OZSDUG and on the three-months anniversary of the 8.0 cessation announcement, I found the ideas generated to be, um, well, psychic.

“Idolize the Flying Librarians of OZ!”

“Vigorously shake the perpetual beta using the OPAC.”

“Raise a toast to SirsiDynix and then claim it isn’t all about technology.”

“Write a song about a perpetual beta and post it to your blog.”

“Mashup SFX links and retire to a safe distance.”

It seems like a safe thing to do, for someone who wonders if ILS products generally are always in beta, or the librarians who are testing them, are always in beta-mode.

Thanks for some great stress relief, Davey!

�

The Opening Keynote Session for CODI 2006 began with CODI Vice President Colleen Medling, CODI President Eileen Kontrovitz, SirsiDynix CEO Patrick Sommers, SirsiDynix CTO Talin Bingham, and SirsiDynix Vice-President for Product Development, Berit Nelson. Gone was the glitz of previous openings, with a more cautious and conservative presentation by all parties. At opening session, it was announced that 788 attendees were present.

Using the mantra of Change, Change, Change, to make things better, Pat gave an overall summary of the state of the company, including

Even if you’re on the right track, you’ll get run over if you just sit there—Will Rogers

Prior to the merger of Sirsi and Dynix, the previous internal company analysis was this:

We were not moving fast enough. We couldn’t serve you more effectively; We couldn’t be the world-class company we wanted to be; We lacked the resources to move ahead; We were NOT satisfied. Like many companies, we concluded that a merger of industry leaders would be the right strategic move. SirsiDynix is [now] No. 1 in worldwide installed base and new sales.

But, as Pat put it,

It’s not the change, it’s the speed of change. Nothing actually killed the dinosauers.

Gary Larsen’s famed cartoon of the dinosauer convention with the CEO Leader lecturing them,

The picture’s pretty bleak, ladies & gentlemen, the world’s climates are changing, the mammals are taking over, and we all have a brain about the size of a walnut – Gary Larsen

brought ample comic relief to the group, and to Pat, I think, as did

The key thing to remember when dancing with a gorilla is this: you don’t stop when you get tired. You stop when the gorilla gets tired. – Joe Flower

Read into your own mind who is the gorilla in this scenario :_)

Pat went on further to state,

Our aim is to bring you value as part of a long-term partnership. To be a real partner, we must do straightforward communications, even if we screwed up. What does a partnership mean? Being up front and open with our customers; No over promising; No misrepresentations; Greater attention to detail; [and] Greater dedication to quality and timeliness.

He went on further,

We’re not where we want to be with 8.0. But we must be honest with each other about what it takes to build a new enterprise system…We will not release products that do not meet our standards.

Enter Talin Bingham, a personable and able engineer who has inherited the 8x product line with Jack Blount’s departure. To a certain extent Talin has had to draw a line in the sand and keep his eye on the future. Talin asked, which of the three of these questions is the most important to the customer:

[Is it most important that new releases be] a) Date-Driven?
b) Feature-Driven?
c) SirsiDynix is Quality/Stability/Performance Driven?

While Talin was hoping all of us would say the answer was c), the real answer for the customer is the fourth and largely invisible choice, d) all of the above. It can’t really be one without the others.

Talin followed with descriptions of new regression and testing tools used by the company to further qualify and quantify the testing and validation of releases and processes within the alpha and beta testing. He further elaborated on server specifications, which should be available elsewhere (presumably on the SD Site or within conference handouts).

Berit Nelson followed with an overview of the new tiered support services for migration and stressed that the scripts and migration utilities will not be run by the customer but will be required to be run by SirsiDynix migration specialists for ALL sites. She listed, of primary importance, the goals in beta testing to be the following:

To open the lines of communication, so that we may learn from your experience and expertise; To provide the highest quality software to beta test partners; To collaborate with test partners to improve the products we provide; To ensure that our software meets the diverse needs of libraries; and To ensure that as much real world context has been given to products.

Berit also had the daunting challenge to explain that the company has now shifted what was 8.1 to 8.0 and indicated that “general release” 8.0 will now include the originally announced 8.1-promised functionality, with a new release date for 8.1 being post 2006. This drew pessimism from the attendees, as there’s usually enough confusion about releases and dates without saying the old release has now become the new and the new is now something else.

All in all, a reserved beginning. Again, documents with details should be available on the CODI website or the SirsiDynix website almost immediately if they are not already available.

The overall impression was one of a somewhat somber reality-check. Is the 8x product exciting? Yes. Does it take a tremendous company-wide investment in resources to get it out and make it right? Absolutely? Do the customers believe yet another year’s platitutes about the state of the product and the state of the company, other than knowing that their ability to use it is still in a wait-state of its own? Skeptically, perhaps again somewhat hopefully. We always grow hopefully as customers when we are surrounded by the energy of our peers and company personnel when we attend CODI. Will the energy be enough to sustain us once we go back home?

As an attendee of the director’s workshop sessions later in the day so elegantly stated in response to a company comment,

Promise? Do you Promise?

I believe the first ever podcasts of CODI presentations are now posted at Kelli’s blog.

If you’re new to podcasts and need to know how to download and listen to them, visit LPL’s podcast info page.

If you’re already familiar with podcasts, here’s the podcast feed:
http://feeds.feedburner.com/kellistaley/podcast

Thank you to all who gave their permission for this!

Industrious members of the Rolling Prairie Library System have posted a report to their constitutents on the recent CODI conference sessions — if you haven’t found a blogged report yet on the session you missed, check out RPLS/LLSAP User Group CODI Report, complete with photos. Additionally, a reminder: if your regional user group has a web site, posted minutes, or has meeting announcements, please post to the Gordian Knot, or let Luke or I know about your regional news. As you can see, we will be able to track much of this on the blog, but our information is only as good as what we receive. :_)

Thanks again to my colleague Kathleen Whitsitt for offering the following notes from the session on Authority Maintenance in Horizon that was presented by Don Carlin, Janet Fox, and Jeffie Stones last Wednesday afternoon:

Don Carlin of Austin Peay State University gave a presentation of how they do automated authority control. He noted that some good times to implement automated authority processing is:

• When adding new libraries in a consortia
• When migrating to a new database
• When you have some money

Janet Fox gave a presentation updating their automated processing at University of Chicago. They have 5.5 million bib records, and 10 million authority records (1.5 million LC or MESH, plus sys gen). They now send out their bib records quarterly to Backstage Library Works. She went over her review process for each quarterly processing including:

• Mistrake log
• Bib_auth_invalid_audit
• Vendor Reports

Jeffie Stones gave a brief description of consulting services available through SirsiDynix for this type of vendor service and contact information. She would not discuss prices. This depends on how much planning and other things they do, which can vary greatly from situation to situation.

After presentations were given, questions were entertained regarding:

• Types of records sent for processing
• Handling of e-resources
• Implementation of local tag preservation in auth records in Horizon 8.0

Thanks to my colleague Kathleen Whitsitt for allowing me to post her notes from Jeffie Stones’ session on Horizon 8.0 Cataloging from last Tuesday:

• Utilities will be incorporated into the cataloging module: Marcin, Marcout, Killitem, Killbib, Killauth should no longer be necessary (don’t know about OCLC holding deletion aspect yet)
• Will support multiple marc formats in same database
• Will convert to/from Dublin Core, Marc21, MarcXML, MODS, etc. on import.
• Consortia can share/not share bib/auth/community records in any combination based on security/access level (but no example or explanation given of how this is supposed to work)
• There can be tag ownership
• Better organized, everything in one place: indexes, marcmaps, import options, match points, tag action will be accessible within the module where they pertain.
• More options for defining matching and overlay of bibs. Can view/change import profile. Can preview records in a file. Can import a portion of a file. Will be able to schedule imports. Can create default items and use item templates.
• More Marc friendly. For instance, in tag action can: add/delete/change subfields, indicator values, set conditions, create/delete $949, link match points.
• Can create lists of imports or searches of marc records (bib, auth, holdings), access individual records from lists, and do batch editing/deleting for marc records. Deleted records will be recoverable up to a point, then removed. Will be able to batch add/edit/delete in item editor, and place batch requests. Will be able to use tag action in Batch Bib Editor to change $65X $x to $v.
• NACO normalization is no longer an option, but standard in authority record import.
• More options, and conditional options in Syntax and Validation. Will be able to employ jurisdictional author mapping $11X.
• Is also supposed to incorporate URL checking, however, Jeffie did not cover this.

Dave Pattern contributes even more marvelously thorough posts on the following sessions:

• “Insights into Web Reporter and NarrowCast”
• “Tailored Just for U: uPortal Customization for Academics”
• “Planning for Hardware: It Doesn’t Have to be Hard”

Kelly Staley has put up her extensive notes from:

• Jack Blount’s 8.0 “SuperSession”, covering Horizon 8.0 Architecture, Planning & Hardware
• “Horizon 7.3 and 7.4 Circulation Tips and Tricks”
• “Creating a Customized HIP for Reference Staff”

This will be a very long post, and also a lot closer to rough transcribed notes than the other posts you’ve seen in our blog coverage. We wanted to make sure this session was thoroughly and accurately covered for the benefit of those that couldn’t be there, and because it was specifically requested by commenters on this blog. Both suzyq and I took notes at this session, and we folded our notes together to fill in areas each other had missed; once we had done that, I also emailed those notes to Jim Wilson, and he was kind enough to take time to look through them and add a few comments and clarifications. What you will see, however, is a very honest, unvarnished, frank discussion. For that, I want to give a lot of credit to those who were up front: moderator Barbara Lowry, and Jim Wilson and Steve Orton from SirsiDynix. So without further ado…

Barbara Lowry, Moderator: We are all in this together. These are our brothers and sisters. This is not a gripe session — it is an opportunity to deal with issues and communicate with each other.

Jim Wilson: “We screwed up.” Previous conversations with Steve Nielsen, company not aware of additional 4.1 problems until users arrived at conference.

Question from floor: What changes are you doing to avoid screwing up in the future?
Jim: 4.0 first tested out on a dozen libraries, went fine; then opened to world, put another 25 on. Fell under the floor, we stopped it. We will stop it if we see something like this again.

Jim: For you who are on 4.x, has your level of contact with the company increased?
Answers from floor: Some yes, some no. (Some indicated it had only increased as initiated by customers.)

Jim: For those still on 4.01, I will not take you a little farther and drop you into another hole. To those that are on 4.01 limping along, we’re going to fix the 4.1 customers first, so that we can get 4.1 stable and then bring 4.01 customers up to 4.1. The first ones that will upgrade to 4.1 are those who are still on 4.01. We will not release 4.1 until the 25+12 are stable and comfortable.

Comment from floor: 4.01 crashed; documentation was incorrect. Support was helpful, but documentation was incorrect. Was the documentation tested?
Jim: Yes, but there were flaws in beta process, and changes have been made to the beta process to avoid this in the future.

Comment from floor: No utilities for 7.4 listed on the website, including killbib marcin/marcout.
Jim: 7.4 was not a huge horizon upgrade, didn’t change much at all. 7.3 utilities, 99% will still work, but it is “our” [company] responsibility to test the utilities, list them, and update the website and make them available for each release.

Question from floor: Can you tell us what the changes to the beta test process are?
Jim: More time with testers in an ASP model of the release before they get it on site. We start training them in more detail before they start. Train in terminology, flow, showing them interpretations of this table of that piece of code. This is how 8.0 beta is being done today. This is being done in 8.0 beta testing today. All are in an ASP environment. This week, several of them will install a version in their site for backroom environment testing.

Question from floor: Doesn’t beta testing imply that the code is there for them to test? It appears that the code is being developed in the 8.0 model as they test.
Jim: to the company, Alpha means “company testers”. Beta means “outside testers”. The code was given to “betas” during this first phase in order that they can actively evaluate during a pliable stage, testing the code while it is written. These testers will take it to production, then the product will be given to early-release candidates.

Question from floor: How long will beta testing occur with sites in production?
Jim: Greater than 30 days. It depends on the testing for a real time frame. We will test as long as is needed.

Jim: If you hear a rumor, don’t spread it. Email Jim Wilson (jim dot wilson at sirsidynix dot com). Call Jim Wilson.

Comment from floor: People at 4.01 / 4.1 felt isolated; it seemed that front-line support personnel were learning the product along with the customers.
Jim: This has changed with 8.0. Trainers are already trained and being trained as the product progresses on 8.0 Support staff are already being trained on 8.0 When it is rolled out, support staff and trainers should already be trained and experienced on it. We are trying to train the whole company, so that when 8.0 does go out, you aren’t experiencing it the first time with support staff who haven’t seen it.

Comment from floor: 4.0 required login use name and password, to the consternation of libraries and patrons. 4.01, same. 4.1 same?
Steve Orton: 4.1 can use barcode and pin or login and password, but not both.

Steve: We appreciate those who suffered through our training. We dumped it and rewrote it. If you have other training comments, please let me Steve Olson or Jeff Olson know. The problem was, if you wanted to use with barcode and pin, it didn’t tie in with Horizon.
Comment from floor: The “other Mr. Wilson” [Jeromy] will synchronize either barcode and pin or username or password.
Steve: It is an either/or for this and this is not fixed in 4.1
Comment from floor: This needs to be a training/setup question which is given attention.

Comment from floor: If you move to 4.1, don’t delete your 1.46 [sic - should be 1.42?] java. 4.1 uses both parts of 1.46 [sic] java and 1.5 java.

Comment from floor: 4.1 still has problems with browse search. Sorting by publication dates. It is not a mass indexer problem. Company is building and doing the loading for the sites, including the mass indexing.

Comment from floor: Lost subject index when company did mass indexing. Will company put it back?
Jim: We are dealing with different issues with each customer. The answer is YES.

Comment from floor: Cannot turn off tracking of patron history. (patrons in their preferences can choose to keep circ history).
Jim: Should come turned off and be patrons’ choice to turn on.
Steve: Location table should have 3 options: tracking on, tracking off, or let patron decide.
Floor: other sites have confirmed that location table is ticked correctly, and it still defaults to keeping track of patron history even if you try to turn it off.

Comment from floor: Bug with PC Reliance as well and patron history tracking.

Question from floor: From the business process standpoint, who makes the decision about limited pre-release vs. general production release?
Jim: Steve Nielsen decides the showstoppers in consultation with his product managers. In 4.1, product manager is Jeromy Wilson. If a beta has an issue, see Jeromy; 4.x product managers should get the information as they speak with betas every week.

Question from floor: how many betas participated in the beta test?
Jim: 10-12 betas per release on average. Then we try to expand it to pre-release sites as a release candidate. Then it will go general.

Comment from floor: Many SAs had performance benchmarks and targets for successful installation and timeframes for other projects within their libraries. These people have been penalized by their institutions because of the faulty products. 4.0 was a release candidate back in April. This has been a very long time in not having anything to show. We are in a stage of disillusion. Our staff attribute every human error they find now to a “Horizon” or “HIP” error. Where are the resources needed to get these 36-38 sites on track with a useable product? Our public are calling, cursing, political entities spending money are concerned; additional local hoops are being imposed on sites now, limiting their ability in some cases now to do incremental build updates.

Comment from floor: Build 56 browse title problems. Browse scoping with consortia and limits still not working;

Comment from floor: We are working with support people only. Jack came on site, which helped with director. But he is not an engineer per se. Your support people are also frustrated. We are still not in conversation with an engineer.
Jim: we are putting the engineer side by side with the person in support. We need to fix your problem. Betas particularly, to get these problems solved.

Question from floor: Why did some libraries get things fixed and others did not?
Jim: Standard procedure for contacts (engineers) to be side by side with support. All will get fixed eventually. Somebody has to be 1st and somebody 2nd. We cannot solve all issues at the same time.

Question from floor: What’s the scoop with Macs and IE 5.5 with the login/password issue? Will the problems on Mac platform be corrected?
Jim: Dynix is committed to supporting IE on Macs.

Comment from floor: I want to thank Jim for the guys who came out [to Howard]. This was a valuable lesson for them. They learned a lot that they can take back.
Jim: Thank you, but shame on us. Some of the libraries that we deal with are in the same town. We should not have had to fly across the country to learn those lessons. We may *not* have a site like you in town. We don’t get or let our product managers out of the office as much as we need to. At the minimum we ought to be able to go over to the ones in our neighborhood and see them in action.

Question from floor: when do you expect to see light at end of tunnel?
Jim: We are working on these logs as we are sitting here. We have people back in the office as we speak who are working as we speak. We feel we are within days or hours of solving these problems.

Question from floor: [Australia/NZ customer] we spent 15 weeks testing. When the logs were passed from Australia to Provo, 1/3 of the issues were quickly fixed; 1/3 of the issues were bugs and things they had not even seen or had reported by US sites. What can you do to incorporate these international issues more comprehensively?
Jim: We are training world wide and improving our tracking methods.
Jim: Do not even thing of going to 4.0. or 4.01. Put it out of your mind. It is possible that you will not see 4.1 or above. [transcription note: Jim wore a shirt covered with HIP 9.0 buttons]

Comment from floor: But there’s a communications problem not just between the company and the beta sites, but between the company and the rest of the customer base. Sales is giving us a different story. How can we be getting straight, direct status information for those who are not here, and next week, and next month?
Jim: At this time, I can’t tell you. If you are not on 4X wait until release and then listen to the list serve. We as a company will try to be more proactive.

Question from floor: Where is the communication? Communication from the betas and group who took an announced general release product and got stuck? Where is your communication with the people who are not at this conference? Where is your communication with directors and others who are paying maintenance?
Comment from floor: We closed the lists, yet you do not use them.
Question from floor: Where is the bug list?
Jim: We will not publish [bug lists for beta products]…
Question from floor: Will you publish what the test and test results of the betas are? Will you publish what the issues are? If there are problems with betas and testing it has to be because the functional specification requirements are somehow flawed.
Jim: Testing procedures have to come from extensive requirements lists. Testing procedures have changed drastically for 8.0/8.1

Jim: let me stress, there is a WHOLE LOT OF DIFFERENCE WITH 8.0 testing than with 7.4 and 4X.

Question from floor: Beta testing: public library should not be testing reserve book room, libraries should be required to test what they know and what they are functional with. Do you need a bigger pool?
Jim: Pool of 10-12 betas of all types is reasonable. Pre-release candidates expand the “ripples” accordingly.

Question from floor: Processor in initial documentation didn’t exist on the unix side; did you target the release on what you wanted to build it on before you built it?
Jim: Yes.

Comment from floor: When moving from beta to “early release candidate,” “known bugs” need to be listed at that stage so that customers can make educated decisions about moving.
Jim: yes, this is true and this list of bugs could possibly be made available, but only to the testing sites. We are not going to rehash all we have fixed for everyone. No Need.
Comment from floor: There is still need for the bug database, or the delta document.

Question from floor: is a test database or test server required for implementation of new software? Implementation > delivery and implementation > are development servers being required? Do we have to backroom everything in order to load a release? One of the problems is that we have backups of our database server, but not necessarily the HIP server.
Jim: We are training our trainers, our support, or project coordinators so they are trained in advance. Production environment, 90% of the customers have a backroom database Question from floor: implementation teams didn’t even recommended it, didn’t even suggest it. Is it required? Sounds like it. If I told them that, they said it would cost. Jim: Our intent is if we help you upgrade on your test system, you will then do the upgrade on the production server. If we do it twice, we charge you. We intend for our releases to be solid enough that you should not have to do this. But if you want to practice and learn on a test database, we expect you to do the production database by yourself.
Comment from floor: it was never suggested that I could do the 2nd one by myself…
Jim: If you want help doing your test server, fine. If you want help doing the production server AND the test server, then I am going to charge you.

[note: The training database is not the backup of the HIP server].

Comment from floor: As customers, we recognize we have responsibilities.
Jim: The company, the implementation team should have suggested that you do this.

Comment from floor: suggestion for pre-release candidates, includes a checklist and disclaimer similar to the beta one.
Jim: I will talk to implementation team should need to suggest and support this.

Comment from floor: when 4.0 released, I attempted to install on background environment. It didn’t work. It is critical also that you release more accurate specifications for test server environments to get accurate sizing into the server specs to minimially use test servers. We don’t need a honker, we need a test environment.

Question from floor: LDAP… LDAP is in 4.1, I can use the same login for HIP as I do for LDAP server, so I don’t remember two logins.
Steve: It is not today [4.1] a single sign-on.

Representing the best spirit of cooperation and peer assistance that CODI is frequently known for, Julio Campa, Miami Dade Public Library System, presented a session on Tuesday on making your own version of HIP in Spanish. We welcome Julio to CODI, and thank him not only for his presentation, but his offer of assistance to any library configuring their ipac, HIP, or C/HIP Portal’s Spanish interface. The Spanish interface currently comes with 3.x or higher. However, manhy of us realized that any local customization or items not on the default profile do not translate unless we edit the strings and provide the translations ourselves. Some libraries are blessed with native speakers in their libraries, but many of us do not. The only thing worse than not having a Spanish profile is having a Spanish profile configured by a systems administrator who thinks she knows Spanish but her college French resurfaces! (I resemble this remark!)

To this end, Julio is a blessing to all of us, and has provided not only step by step instructions, on how to do the setup and modification of the strings, profiles, and variables, but he continues to offer his direct translation of various library catalog phrases for the upcoming releases. While Julio is a valuable employee to Miami Dade with a day job that keeps him quite busy, we thank him (and his director) for this gracious offer in the spirit of cooperation and giving.

It looks like while the HIP 4.1 and higher releases will come with a full Spanish version, any customizations will again need to be re-translated by hand for local tags, buttons, and strings when the library has chosen to modify the default. Therefore, Julio is guaranteed job security in his unofficial post of “translator laureate” to the user group for many years to come. Thanks, Julio, for a great presentation, and your willingness to lend a hand to all of us!

“We screwed up.” — Jim Wilson, on What Really Happened with 4.0?