Access Conference 2018 Day 1 Afternoon Sessions #AccessYHM
Data Migration to Open Journal System (OJS) Using R
You Young Lee
Worked with scholarly communication librarian to move from legacy system to OJS. Migration using R programming language. Wanted to migrate nursing journal Aporia. Internal system didn't support editor peer review, no user friendly interface to upload articles, typical workflow was to receive content and copy and paste metadata using templates and uploading to server. Since OJS is available, time to move on. Approximately 32 issues, didn't want to manually copy and paste. New issue will have updated location.
Project goal: how to convert metadata in HTML into XML. Project tool and packages: Use Studio instead of R, provides console and editor, debugging tool and workspace management. Easy to manage datasets and scripts all in one place. Rcrawler, data.table, dplyr, XML, stringr. Challenge 1: how download data from the website: crawler parses whole website and extracts all data with a single command line. HTML classes. RCrawler commands. Outputs 3 things
Data variable outputted had 5 different types of element- list of 5 - title, abstract. authors, etc. Convert this list to data frame. Data cleaning. Now data modification - how to populate missing values like volume number, issue number, and year. mutate function in dplyr adds new variables that are functions of existing variables.
How create OJS XML file from data frame? Developed a custom function to transform.
Learned R for statistics courses for data analysis. Similar to Python or scripting languages.
GAME ON!!!: Building Interactive Educational Fiction or FAILURE: Games are Hard?
Ruby Warren
Interactive fiction because people learn better when they have an avatar to connect with. Python based visual . Attention Relevance Confidence Satisfaction (ARCS) model of motivation, also don't work with game engine initially chosen, so had to change prototypes.
Pitfall 1: Don't do everything in the wrong order. Need project routine then timeline. Don't unintentionally use waterfall method of software design. Set learning goals before everything else. let learning goals and learning type inform engine choice. Project outline THEN timeline. Keep in touch with reality when doing timeline. She assumed vast majority of time would be construction, underestimated time for tool familiarity, underestimate learning design time
Pitfall 2: Double time estimates. Then triple them. Budget adequate time for: learning outcome design, engine familiarity, customization scripting, everything being on fire.
Why: forgot to consider user motivations in using the product; developing appeals takes time; didn't plan for testing during development. Bad from UX design standpoint. First HTML5 and Javascript. Don't forget the user - needs to be browser based, easy to load, doesn't take forever. Plan for apathy, plain for testing/trials. Don't underestimate appeal.
Archive Your ILS: How to Keep a Copy of Your Old ILS when you Migrate to a New ILS
Calvin Mah
Major ILS migration - moved from III Millennium (20 years) and moved to Ex Libris' Alma/Primo. During migration, concerned that Ex Libris wouldn't migrate everything. For majority of the records (bib and patron) fine, but the missing things like order records would not get migrated, required collections staff manually recreate them in new system, needed to do a lot of data cleanup, and order records had a long history of notes and correspondence where staff journal experiences with vendor that wouldn't be in new system. Patron fines - only fine amount would be migrated. In III can see fines and how separated by cause of fine. Ex Libris only migrated total fine amount, other information is lost. Needed current unpaid fines to Alma. Checkin cards for serials were also not to be migrated. Another thing some staff had issue with: they didn't trust the process would have exactly what there was before. They wanted ability to go back to old system and compare. A month of overlap due to licenses, but after that can't legally run Millennium. Solution: print out records? NO, but people were thinking of doing this.
Biult an archival copy - actual database like a mini ILS. Designed in such a way that it's a web app simple to use, take advantage of data extracts that had to do anyway regularly for Ex Libris migration. One copy to Ex Libris, another copy to load into archive. In archive can do simple bib record searches - get some metadata with Millennium specific data, MARC record, attached item record. Additionally patron records are searchable. Single cup sum was migrated, but added fine details - Millennium didn't have this as exportable so he got a list of all patron barcodes with fines and scripted a fake login to each patron to view library fines and screenscraped all the records and added ot the archived data.
Challenges: no one wanted to use it, wanted live Millennium system, when old system was gone then realized items weree gone they needed, too late. They wanted it to behave just like Millennium did. Another weird thing was that when you searched for The New Yorker they wanted search to respect second indicator non filing skip (?) and to find "New Yorker." Privacy and security - patron data loaded into here is running on library servers, so only people only normally able to view this data should have this access. Archive is running on something (?) slow, so added sole indexing to speed up searching. Info in archive is relevant but is getting stale, it was a snapshot at migration, eventually need ot move folks from archival to actual--staff reliance.
learned: handy and lots of use when old system went offline. the loans counter was used a lot to look at fines details. Legacy of archived patron data - load fines in to Alma patron notes and delete from patron archive for security and privacy risk.
Contract Work Means Work for You Too: How to Make Your Project Sustainable After the Contract is Over
Bobbi Fox
What is contract work
preplanning
Defining contact
While contractor is on the job
Getting closure
Definitions - contract work is scoped out limited work (project) performed by someone for a given amount of hours or time. Contractors - person or group selected to do the work. She is a developer whose decades of experience include being a contractor, team lead with contractors on team, a maintainer/enhancer of contracted-for code. Also, IANAL.
Need to try to ensure sustainability of project once contact is over. Minimizing frustration - no such ting as a turnkey application. It always takes more time than you expect. Try to minimize ongoing support work. project is more than delivered product - someone has to support it going forward. Need to build a bridge to your IT group.
- PrePlanning: Talk to IT Group: What technologies does IT support? Programming languages, hardware, system configs, etc. Ex - all apps go through load balancer at Harvard which means that app only sees load balancer IP and not user's. Are the operations people willing able to support project? Are there resources for development/upgrades? Who provides end user support? Who "owns" the project going forward?
Defining the contract: think about minimum viable product. What technologies have already addressed it. technologies (language, hardware, operating system, deployment setup/ system configurations). Accessiblity not just WCAG 2.0, but whether will be accessible to those without latest and greatest smartphone or browser. Version control: pick a system and make the contractor use it during development- GitHub, other options. Also semantic versioning and tag release. Configuration: system should have QA version as well as production version. Configuration. Tests - tests of the way the software should work. Logging - meaningful log messages for errors.
Documentation - installation/Deployment don't assume IT group knows how to deploy. technical documentation within code but useful (why not how). User oriented documentation. What does end user support group need to know? All should be a part of each delivered version, not just at the end.
What you should expect from contractor - good knowledge of an WILLINGNESS TO USE the languages the work is to be done in. If using open source or adding to already built application, an understanding that these applications may ALSO get version updates, and coding accordingly. Understanding of customization hooks that already exist. Willingness to expose not-yet-perfect code for review, if appropriate.
While contractor is on the job:
periodic code reviews
etc.
Things can go wrong - early termination (quit or fire) - no continuity. equally dire if they're just not doing the job you hired them for. Contract management is not for the conflict averse. Running out of time/money - either way, hopefully have minimum viable product, but finish project or throw up hands and call it a day.
Getting closure.
When the Digital Divides Us: Reconciling Emerging and Emerged Technologies in Libraries
Monica Rettig, Gillian Byrne, Krista Godfrey, and Rebecca Laroque
Will be kind intent. Libraries are spreading more than shifting - spreading, not *moving*. Evidence for continuing needs of basic things like computers, working wifi, etc and training to help patrons to use it remains strong. The creates disconnects - IT resources required for boutique nature of emerging technologies, care of patron experience vs print management systems.
@weelibrarian @gillmbyrne @monicarettig and @rlarocque paneling at #AccessYHM
Core Topics in the Web Content Accessibility Guidelines
Mark Weller
WCAG 2.0 (Web COntent Accessibility Guidelines). Guidelines to help developers develop web content hat could be used by people with disabilities. 2.0 was adopted in 2008. Jurisdictions around the world have made these requirements (government of Canada, more). In the US this year WCAG was adopted to rehabilitation act of 1978, and ADA is using it as a standard, used as standard for higher education. EU directive on WCAG 2.0.
Ellcessor, Ellen. (2010) - ??
Global Public Inclusive Infrastructure
4 official documents: WCAG 2.0 (31 pages), Understanding WCAG 2.0 (200 pages), techniques for WCAG 2.0 (600 pages), How to Meet WCAG 2.0 (reference guide). Style: tone is formal, vocabulary precise, treatment of topics is comprehensive, organization intricate. The 31 pager isn't technical in nature, by design. Referring to specific technologies would make it outdated quickly. Abstractions themselves are based on a usability framework that has 3 levels - level 1 principle of usability: person must be able to perceive, operate it, understand it, needs to be robust enough . level 2 - 12 guidelines. Level 3 - 61 success criteria organized in 3 levels. Also defines 5 requirements of conformance. Complex topic needing international dialogue.
moving from abstract into technical layers of building the web content. Reliable techniques: 383 reliable techniques. 150 apply to all technology. 233 technology-specific techniques (HTML, PDF, Flash). Each technique comes with a testing procedure, need to know how to interpret test results. Stable - since 2014, only 8 added. As circumstances change, need to add to collection of reliable techniques. Also detected 75 common problems as well as techniques to test if problems exist or not. Just because passes one test, doesn't mean content passes all tests.
Practical Linked Data Implementation: Trail of the Caribou
Heather Pretty
Data collected by her husband, now held by Trail of the Caribou - all stories are important. Newfoundlanders in WWI. 6400 served, 1304 KIA, 5 monuments to major battles of regiment. 7500 other Newfoundlanders served another militaries. Some of the types of data collected for each soldier - name, regiment number, unit, hometown, trade, etc. Held across multiple spreadsheets, messy. Linked data offers ability to answer complex questions across multiple sources. Tim Berners Lee 4 rules of linked data 2006. URIs into RDF triples. Each object points to a different source.
Linked data and the semantic web are about teaching computers individual terms using URIs.
Robert Chavez courses through Library Juice Academy
Learning SPARQL 2nd ed by DuCharme
Mastering Structures data in the semantic web
linked data structured data on the web
Look at bib frame, transition catalog and get subfield $0 into bib record
How long was learning curve? Yearlong sabbatical.
You Young Lee
Worked with scholarly communication librarian to move from legacy system to OJS. Migration using R programming language. Wanted to migrate nursing journal Aporia. Internal system didn't support editor peer review, no user friendly interface to upload articles, typical workflow was to receive content and copy and paste metadata using templates and uploading to server. Since OJS is available, time to move on. Approximately 32 issues, didn't want to manually copy and paste. New issue will have updated location.
Project goal: how to convert metadata in HTML into XML. Project tool and packages: Use Studio instead of R, provides console and editor, debugging tool and workspace management. Easy to manage datasets and scripts all in one place. Rcrawler, data.table, dplyr, XML, stringr. Challenge 1: how download data from the website: crawler parses whole website and extracts all data with a single command line. HTML classes. RCrawler commands. Outputs 3 things
Data variable outputted had 5 different types of element- list of 5 - title, abstract. authors, etc. Convert this list to data frame. Data cleaning. Now data modification - how to populate missing values like volume number, issue number, and year. mutate function in dplyr adds new variables that are functions of existing variables.
How create OJS XML file from data frame? Developed a custom function to transform.
Learned R for statistics courses for data analysis. Similar to Python or scripting languages.
GAME ON!!!: Building Interactive Educational Fiction or FAILURE: Games are Hard?
Ruby Warren
Interactive fiction because people learn better when they have an avatar to connect with. Python based visual . Attention Relevance Confidence Satisfaction (ARCS) model of motivation, also don't work with game engine initially chosen, so had to change prototypes.
Pitfall 1: Don't do everything in the wrong order. Need project routine then timeline. Don't unintentionally use waterfall method of software design. Set learning goals before everything else. let learning goals and learning type inform engine choice. Project outline THEN timeline. Keep in touch with reality when doing timeline. She assumed vast majority of time would be construction, underestimated time for tool familiarity, underestimate learning design time
Pitfall 2: Double time estimates. Then triple them. Budget adequate time for: learning outcome design, engine familiarity, customization scripting, everything being on fire.
Why: forgot to consider user motivations in using the product; developing appeals takes time; didn't plan for testing during development. Bad from UX design standpoint. First HTML5 and Javascript. Don't forget the user - needs to be browser based, easy to load, doesn't take forever. Plan for apathy, plain for testing/trials. Don't underestimate appeal.
Archive Your ILS: How to Keep a Copy of Your Old ILS when you Migrate to a New ILS
Calvin Mah
Major ILS migration - moved from III Millennium (20 years) and moved to Ex Libris' Alma/Primo. During migration, concerned that Ex Libris wouldn't migrate everything. For majority of the records (bib and patron) fine, but the missing things like order records would not get migrated, required collections staff manually recreate them in new system, needed to do a lot of data cleanup, and order records had a long history of notes and correspondence where staff journal experiences with vendor that wouldn't be in new system. Patron fines - only fine amount would be migrated. In III can see fines and how separated by cause of fine. Ex Libris only migrated total fine amount, other information is lost. Needed current unpaid fines to Alma. Checkin cards for serials were also not to be migrated. Another thing some staff had issue with: they didn't trust the process would have exactly what there was before. They wanted ability to go back to old system and compare. A month of overlap due to licenses, but after that can't legally run Millennium. Solution: print out records? NO, but people were thinking of doing this.
Biult an archival copy - actual database like a mini ILS. Designed in such a way that it's a web app simple to use, take advantage of data extracts that had to do anyway regularly for Ex Libris migration. One copy to Ex Libris, another copy to load into archive. In archive can do simple bib record searches - get some metadata with Millennium specific data, MARC record, attached item record. Additionally patron records are searchable. Single cup sum was migrated, but added fine details - Millennium didn't have this as exportable so he got a list of all patron barcodes with fines and scripted a fake login to each patron to view library fines and screenscraped all the records and added ot the archived data.
Challenges: no one wanted to use it, wanted live Millennium system, when old system was gone then realized items weree gone they needed, too late. They wanted it to behave just like Millennium did. Another weird thing was that when you searched for The New Yorker they wanted search to respect second indicator non filing skip (?) and to find "New Yorker." Privacy and security - patron data loaded into here is running on library servers, so only people only normally able to view this data should have this access. Archive is running on something (?) slow, so added sole indexing to speed up searching. Info in archive is relevant but is getting stale, it was a snapshot at migration, eventually need ot move folks from archival to actual--staff reliance.
learned: handy and lots of use when old system went offline. the loans counter was used a lot to look at fines details. Legacy of archived patron data - load fines in to Alma patron notes and delete from patron archive for security and privacy risk.
Contract Work Means Work for You Too: How to Make Your Project Sustainable After the Contract is Over
Bobbi Fox
What is contract work
preplanning
Defining contact
While contractor is on the job
Getting closure
Definitions - contract work is scoped out limited work (project) performed by someone for a given amount of hours or time. Contractors - person or group selected to do the work. She is a developer whose decades of experience include being a contractor, team lead with contractors on team, a maintainer/enhancer of contracted-for code. Also, IANAL.
Need to try to ensure sustainability of project once contact is over. Minimizing frustration - no such ting as a turnkey application. It always takes more time than you expect. Try to minimize ongoing support work. project is more than delivered product - someone has to support it going forward. Need to build a bridge to your IT group.
- PrePlanning: Talk to IT Group: What technologies does IT support? Programming languages, hardware, system configs, etc. Ex - all apps go through load balancer at Harvard which means that app only sees load balancer IP and not user's. Are the operations people willing able to support project? Are there resources for development/upgrades? Who provides end user support? Who "owns" the project going forward?
Defining the contract: think about minimum viable product. What technologies have already addressed it. technologies (language, hardware, operating system, deployment setup/ system configurations). Accessiblity not just WCAG 2.0, but whether will be accessible to those without latest and greatest smartphone or browser. Version control: pick a system and make the contractor use it during development- GitHub, other options. Also semantic versioning and tag release. Configuration: system should have QA version as well as production version. Configuration. Tests - tests of the way the software should work. Logging - meaningful log messages for errors.
Documentation - installation/Deployment don't assume IT group knows how to deploy. technical documentation within code but useful (why not how). User oriented documentation. What does end user support group need to know? All should be a part of each delivered version, not just at the end.
What you should expect from contractor - good knowledge of an WILLINGNESS TO USE the languages the work is to be done in. If using open source or adding to already built application, an understanding that these applications may ALSO get version updates, and coding accordingly. Understanding of customization hooks that already exist. Willingness to expose not-yet-perfect code for review, if appropriate.
While contractor is on the job:
periodic code reviews
etc.
Things can go wrong - early termination (quit or fire) - no continuity. equally dire if they're just not doing the job you hired them for. Contract management is not for the conflict averse. Running out of time/money - either way, hopefully have minimum viable product, but finish project or throw up hands and call it a day.
Getting closure.
When the Digital Divides Us: Reconciling Emerging and Emerged Technologies in Libraries
Monica Rettig, Gillian Byrne, Krista Godfrey, and Rebecca Laroque
Will be kind intent. Libraries are spreading more than shifting - spreading, not *moving*. Evidence for continuing needs of basic things like computers, working wifi, etc and training to help patrons to use it remains strong. The creates disconnects - IT resources required for boutique nature of emerging technologies, care of patron experience vs print management systems.
@weelibrarian @gillmbyrne @monicarettig and @rlarocque paneling at #AccessYHM
Core Topics in the Web Content Accessibility Guidelines
Mark Weller
WCAG 2.0 (Web COntent Accessibility Guidelines). Guidelines to help developers develop web content hat could be used by people with disabilities. 2.0 was adopted in 2008. Jurisdictions around the world have made these requirements (government of Canada, more). In the US this year WCAG was adopted to rehabilitation act of 1978, and ADA is using it as a standard, used as standard for higher education. EU directive on WCAG 2.0.
Ellcessor, Ellen. (2010) - ??
Global Public Inclusive Infrastructure
4 official documents: WCAG 2.0 (31 pages), Understanding WCAG 2.0 (200 pages), techniques for WCAG 2.0 (600 pages), How to Meet WCAG 2.0 (reference guide). Style: tone is formal, vocabulary precise, treatment of topics is comprehensive, organization intricate. The 31 pager isn't technical in nature, by design. Referring to specific technologies would make it outdated quickly. Abstractions themselves are based on a usability framework that has 3 levels - level 1 principle of usability: person must be able to perceive, operate it, understand it, needs to be robust enough . level 2 - 12 guidelines. Level 3 - 61 success criteria organized in 3 levels. Also defines 5 requirements of conformance. Complex topic needing international dialogue.
moving from abstract into technical layers of building the web content. Reliable techniques: 383 reliable techniques. 150 apply to all technology. 233 technology-specific techniques (HTML, PDF, Flash). Each technique comes with a testing procedure, need to know how to interpret test results. Stable - since 2014, only 8 added. As circumstances change, need to add to collection of reliable techniques. Also detected 75 common problems as well as techniques to test if problems exist or not. Just because passes one test, doesn't mean content passes all tests.
Practical Linked Data Implementation: Trail of the Caribou
Heather Pretty
Data collected by her husband, now held by Trail of the Caribou - all stories are important. Newfoundlanders in WWI. 6400 served, 1304 KIA, 5 monuments to major battles of regiment. 7500 other Newfoundlanders served another militaries. Some of the types of data collected for each soldier - name, regiment number, unit, hometown, trade, etc. Held across multiple spreadsheets, messy. Linked data offers ability to answer complex questions across multiple sources. Tim Berners Lee 4 rules of linked data 2006. URIs into RDF triples. Each object points to a different source.
Linked data and the semantic web are about teaching computers individual terms using URIs.
Robert Chavez courses through Library Juice Academy
Learning SPARQL 2nd ed by DuCharme
Mastering Structures data in the semantic web
linked data structured data on the web
Look at bib frame, transition catalog and get subfield $0 into bib record
How long was learning curve? Yearlong sabbatical.
Comments