BHL Survey Results

Back to BHL Survey questions (English) | BHL Survey Home

Preliminary results

http://www.planetposter.de/bhl/wien2010.htm -- Data compiled by Fracisco Welter-Schultes

Analysis of Q9 & Q10 Free Text answers (English)

Overall Themes

1. The overwhelming, number one comment/complaint from users is Search. Users are frustrated with our search, and particularly want to be able to do “fuzzy” searching. Below are some other search related comments that came up regularly:
   • Ability to do Boolean search
   • Ability to do keyword search
   • Ability to “search inside book”
   • Ability to combine search criteria
   • Ability to search specifically for illustrations
   • Maintain search string in search box even after search results come up
2. The number two complaint that I saw was that there are so many gaps in serial runs. Users want these gaps filled
3. The number three complaint was with image quality
4. Be able to print/download one page at a time/print on demand features that don’t require generating your own PDF (selecting pages to download)
5. Feature such as “you might be interested in” or feature that shows works related to the work the user accessed
6. Article indexing and searching at the article level
7. Request specific titles
8. Easily switch between OCR text and page images
9. Mechanism to obtain user-submitted corrections, such as to OCR, pagination, enumeration, etc.
10. Ability to filter results by date
11. Ability to bookmark “favorites”
12. Implement more meaningful user-generated PDF file names (such as file name contain title and author)

BHLSurvey.xlsx

Analysis Methodology

Each Free-text answer was categorized according to:

• 0 = Other, never used BHL before, comment not really offering anything substantial
• 1 = Praise
• 2 = Criticism, listing 3 or fewer problems
• 3 = Criticism, listing more than 3 problems

Category 2 assigned the following themes (ea. answer could be assigned up to 3):

1. Articles
2. Citations
3. Communication
4. Content
5. Copy paste
6. Duplication
7. Downld single pg
8. Gap-fills
9. Global BHL
10. Google
11. Help
12. Highlight (ability for UI to highlight search terms within text)
13. Img quality
14. Linking data
15. Metadata
16. OCR
17. Pagination
18. Pdf
19. Printing
20. QA
21. Recent additions
22. Requests
23. Search
24. Search inside
25. Sorting
26. Speed
27. TL2
28. UI

Q9 Results Summarized

Category Break-down

English language survey

Row Labels	Count of Scale
0	47
1	16
2	125
(blank)
Grand Total	188

Multi language survey

Row Labels	Count of Scale
0	37
1	11
2	72
3	2
(blank)
Grand Total	122

Theme Break-down

Note: For all answers in category 2 (categories 0,1,3 not assigned themes). Results below combine both the ENGLISH language only survey + the MULTI language survey to show grand totals by theme.

Theme	Count
articles	13
citations	6
communication	3
content	12
copy paste	1
duplication	1
dwnld single pg	4
google	18
highlight	5
img quality	6
linking data	3
metadata	9
OCR	2
pdf	11
printing	1
QA	1
recent additions	1
search	98
search inside	15
sorting	10
speed	11
UI	16

Q10 Results Summarized

Category Break-down

English language survey

Scale	Count of Scale
0	13
1	71
2	124
3	11
(blank)
Grand Total	219

Multi language survey

Row Labels	Count of Scale
0	17
1	43
2	43
3	3
(blank)
Grand Total	106

Theme Break-down

Note: For all answers in category 2 (categories 0,1,3 not assigned themes). Results below combine both the ENGLISH language only survey + the MULTI language survey to show grand totals by theme.

Theme	Count
Articles	7
communication	8
content	24
copy paste	1
duplication	5
dwnld single pg	4
gap-fills	17
Global BHL	1
Google	2
Help	1
img quality	12
metadata	11
OCR	8
pagination	5
pdf	19
printing	7
QA	3
recent additions	2
requests	12
Search	27
search inside	6
sorting	2
speed	6
TL2	1
UI	18

Freetext answers Q9 and Q10: fr French, de German, sp Spanish, it Italian, po Portuguese

It was not necessary to subdivide Q9 and Q10 - we had to do the analyses twice and people repeated themselves.

Search function:

several articipant asked us generally to improve the search function
9de4 - full text search
9de7, 9it2, 9it6, 9it8 - search by year, format, size
9de10 - refine search methods in Google scholar (but there the return set is not good)
9de20 - need to improve performance
9de21 - likes advanced search function in Google
9de23, 9sp9 - search inside fulltext
9de25 - logical operators and wildcards
9de31 - Google scholar: search backwards
10de3 - ability to search by taxonomic name with wildcard ption, also search by specific name without generic name
9de30, 10sp16 - ability to search for articles in journals
10de18 - ability to search by animal group, family
9sp2 - ability to restrict search by years
9sp14 - speed
9sp15 - Botanicus has a too complicate search function

Return set of results:

9de9 - Google scholar: ability to show all found versions
9de9, 9de14, 9sp7 - noting how many times and where a book has been cited in other publications
9de15 - ability to show all found versions
9de24 - ability to show more volumes of the same series
9de32 - Digitial Library Madrid has good listings of authors and works
10de20 - year and volume number should be shown prominently in the search result
10de25 - ability to filter by new recently scanned works within a results set (= sort by: date of contribution)
9sp6 - preview of title page in results set

Ability to use citations (Google scholar)
9de8
9de36

Create my own virtual library function (Google books)
9de33

"will be scanned"
9sp5 - AnimalBase: ability to see if a work will be scanned (=has a bid on the bidlist), with approximate date when it will be online

tagging by geographical region
10de11

find likewise literature
9de14

links to other databases
9de26
10de15 - example NCBI

Contents

9fr6 - very rare works like in AnimalBase
9de22, 10de23 - more works on paleontology
10de14 - more works on Protozoa, Fungi and Bacteria
9de11, 10fr5, 10de11, 10de12, 10de32 - more modern literature after 1922, particularly in serials
10fr5, 10sp9, 9po2 - filling gaps of serial runs
10de1 - ablity for the system to register if a volume was searched and not found - s that such a volume could automatically placed on the candidate lists for scanning
10de29 - better coordination, deduplicate
10po2 - ability to upload self-scanned literature

File formats

PDF:
9fr1, 10sp10 - download articles of serials
9de6, 9it9 - improve PDF dowload
10de30 - PDFs should be compatible downwards, so that not only the most modern readers should be abe to read the files

OCR
9de13 - ability to correct OCR manually
10de30 - problems of compatibility

Images
10de21 - not only JP2, also TIF, RAW, JPG

scan quality, quality control
10de17 - black on white, not brown on tan
10de21 - better in JSTOR: tan removed, stamps and handwritings are not interesing, OCR result in clean version is better
10de28 - please test completeness and scan quality before releasing a work online
10de31, 9it3 - scan plates with highr resolution and better image quality, details must be visible
9sp15 - very bad quality control in Google books and Gallica (many pages often missing)

online viewer
9fr5 - search inside online viewer
10de27 - online viewer should provide a quick view default option, without brown tan
9sp15 - quick access (AnimalBase works load in to sowly because too large file sizes)
10sp16 - should be a function/button that one can quickly access the title page from somewhere inside the work (hitting the back button is problematic)

full screen function
9de28


Taxonomic names in publications
9de13 - ability to mark taxonomic names in a publication

Current allocations of taxonomic names
9de16

Metadata

page level metadata (plate numbers)
9de28, 10de10, 10de34 - plate numbers

structural metadata (chapters in books)
9de13 - chapters in books
10de10, 10sp7 - volume numbers should be correct and complete (including hefts where necessary)

Other

stable URL
9de13

APIs
10de8 - should be open source and work also in Linux and Mac systems