New Suggestion Form

[richard-jones]

New Suggestion Form

1. Extend the existing data model and provide the documentation
2. Create the new Suggestion form, with client and server-side validation for form content, auto-complete where appropriate, and javascript driven input assistance where suitable.

Note that the Suggestion form has implications for the data model in DOAJ, and while we will update the data model to handle the additional content, this work item does not include propagating those changes to the public-facing user interface(s).

[richard-jones]

Some questions for the data model:

1. Society - is this the same as "provider"? If so, can we continue to use the term "provider" or should we migrate the whole dataset over to "society"?

   DM: no. provider is now platform/host/aggregtor. See below. Society is a new field.

2. Platform/Host/Aggregator - I'm still not totally clear what this is. One of the examples is "Highwire Press"; so this is if there is a service provider who hosts the journal for you? Would "OJS" be another value you might enter here?

   DM: yes, OJS would work here. For example: the journal 'Heart', owned by the British Cardiac Society (society), is published by BMJ Publishing Group (publisher) and made available online via HighWire Press (platform).

3. APCs - at the moment we have author_pays, which is not boolean (unlike the new form); it can contain yes, no, conditional and no information. Should we have a new field for APC, and do we have to migrate author_pays?

   DM: retain as Yes/No, as it is in the new form. Ignore/delete all existing data for conditional and no info.

4. There is a question relating to the BOAI, and then a question about the nature of the license. We could reduce this to one question ("which licence are you using") and infer the BOAI compliance from that. There is space in the data model already for this question, so that's good.

   DM: unfortunately not. There are places in the world where CC licenses aren't valid and we have many publishers who don't user CC licenses. So we need to retain both.

5. There is a question in the form example numbered 41 which has a year placeholder XXXX - what is that supposed to indicate? Is it a fixed year you will decide, or is it dynamic? This question is quite complex, and may benefit from being broken down into two answers: author retains copyright? author retains publishing rights (or similar)

   DM: please replace XXXX with 2015. It was a placeholder. I have now split this into two questions.

[richard-jones]

And some more questions:

1. Deposit Policy - do we need to collect any urls for this, or is just knowing that it is in, eg, Sherpa/Romeo enough

DM: no URLs needed.

1. When we store the full-text formats, I think we should store the mimetypes explicitly (e.g. PDF = application/pdf, HTML = text/html)
2. What is the Open Access statement? Is this different from the URL of a licence the journal might apply?

DM: yes, as above, many publishers do not apply licenses.

1. Does the journal provide download statistics - the instructions ask us to collect a url, but I'm not sure what the URL would be. Presumably, the download stats will be per article?

DM: have changed the wording on the site and made the URL non-compulsory.

1. The questions around re-use/open access/licensing seem like a round-about way of collecting the licence information. I think that the best way to approach this would be just for us to ask them what their licence statement is, and then infer the remaining things (e.g. from the CC licence we can determine BOAI compliance). I think it would be good to have a talk about this in detail in our next meeting

DM: OK but as per above. Not all journals uses licences so we need to try to capture as many angles of this as possible.

1. The existing model has space for OA start volume and issue, but these are not reflected in the suggestion form - is that correct?

DM: this is correct. CL should ignore/delete all existing data in these fields.

1. Alternative title is not included in the suggest form

DM: are CL able to tell me how many of the journals have a value for this field?

1. The active flag is not in the suggest form - but is this something that the administrators would still need to use?

DM: correct

1. Subject classification is not included in the suggest form - I think this is right, but just wanted to check

DM: correct

1. The suggestion description is no longer included in the suggest form - is that correct?

DM: correct

1. The "suggested by owner" field has gone - from recent conversations I suspect this is correct, but just checking.

DM: correct

[richard-jones]

I have created a spreadsheet which starts to map the form to the data model, and also has some notes about how we can aid the user input:

https://docs.google.com/a/cottagelabs.com/spreadsheet/ccc?key=0AsW3CmDjW93kdGlRU21FNFZSRHhPaUk0ZmtUUHdGNnc#gid=0

[dommitchell]

Have updated your comments with our responses. One question for CL: Alternative title is not included in the suggest form DM: are CL able to tell me how many of the journals have a value for this field?

[emanuil-tolev]

One question for CL: Alternative title is not included in the suggest form DM: are CL able to tell me how many of the journals have a value for this field?

Sorry, didn't see this.

9609 out of 12459 journals are missing an alternative title or have an empty one.

[dommitchell]

Is there standard terminology for article submission charges? DM: no, there isn't. Lars suggested just submission charges but in your data model you could probably use whatever you wanted.

Write out some scenarios for questions 39 and 41. DM: 1) an author submits a paper to an open access journal. The journal has a CC-BY license for all content. When the paper is accepted for publication, the publisher asks the author to sign a form transferring all copyright and reproduction rights over to the publisher indefinitely. In this scenario, the author might not choose to send her article to this publisher as any activity by the author after publication would be limited. 2) an author submits a paper to an open access journal. The journal has a CC-BY license for all content. When the paper is accepted for publication, the publisher asks the author to sign a form that grants the publisher the right to reproduce the work for one year but all copyright and permissions stay with the author. This is definitely better for the author in this scenario.

Is it confusing that these 3 are lumped together? Should there be subheadings? DM: I have made changes to the form to reflect the different areas.

Do we need alternative title? DM: yes and I have added it into the form.

Due to these changes, the form numbering is now out of order. Please ignore it. I will correct it if you want me to.

[richard-jones]

For the record, on implementing this form. The questions "From 2015 ..." should all be this year + 1, so next year it becomes "From 2016...." automatically

[dommitchell]

Further comments from my call with Lars: 3) Will data from the article count (question 13) be something that will be part of a journal's published metadata? DM: yes and will be updated every year. Preferably by the publisher. and then approved by the EdTeam.

4) Shall we include CC0, CC-BY-SA as options for license? DM: we will not include CC0 but CC-BY-SA should be included. I have updated the form already but see below... Lars liked your idea of adding the version number but said that it should not be a compulsory question because many publishers may not know which version of the CC license they have. Does that make sense or is this an unnecessary worry?

5) Confirm question 39 with Lars, especially wording "in accordance with". (Currently, CL will be displaying that if CC-BY selected, CC-BY displayed against journal, even if they have something LIKE a CC-BY.). DM: we take onboard of all CL's feedback here and wondered if the following solution might be possible:

/start/ Does your journal use a CC license or an equivalent thereof? If you do not apply a CC license to your content but have something similar, please go to http://creativecommons.org/choose/ and find out which CC license matches your journal. Left column Right column OR CC-BY CC-BY equivalent CC-BY-SA CC-BY-SA equivalent CC-BY-NC etc etc CC-BY-ND /end/

Then on the site, each journal would display the value of the radio button selected.

Could CL go over their idea again for having publishers choose attributes?

Finally:

For the record, on implementing this form. The questions "From 2015 ..." should all be this year + 1, so next year it becomes "From 2016...." automatically

Actually, we followed CLs advice and have removed this from the question altogether.

[dommitchell]

Question: for publishers who have many journals, will bulk upload via a spreadsheet upload option be possible?

[emanuil-tolev]

Sorry, just to clarify: bulk creation of suggestions via a spreadsheet (with ~50 columns)?

[dommitchell]

Yeh, I was thinking of a way that the publisher might be able to provide a csv file with the info that we could ingest straight in to the database, so they didn't have to use the form 150 times

[richard-jones]

I have opened a new issue #106 to cover this possibility.

[richard-jones]

Some comments on the above further feedback

3) Will data from the article count (question 13) be something that will be part of a journal's published metadata? DM: yes and will be updated every year. Preferably by the publisher. and then approved by the EdTeam.

RJ: I think we're talking about slightly different things, so need to clarify. The form says "do you supply article metadata", which I have taken to be a flag (true or false) as to whether the publisher makes stats available. But are you also saying that you want to capture those statistics?

/start/ Does your journal use a CC license or an equivalent thereof? If you do not apply a CC license to your content but have something similar, please go to http://creativecommons.org/choose/ and find out which CC license matches your journal. Left column Right column OR CC-BY CC-BY equivalent CC-BY-SA CC-BY-SA equivalent CC-BY-NC etc etc CC-BY-ND /end/

The things to take into account here are:

• what are we actually going to store in the database
• what are we going to present to the end user in the discovery interface

I'm still not clear, so we should go through this again on Monday.

DM: Could CL go over their idea again for having publishers choose attributes?

I was thinking something along the lines of:

Choose your licence:

• CC-BY
• CC-BY-SA
• CC-BY-NC
• Other:

if "Other", does it require:

Attribution [ ] Share Alike [ ] No Commercial Usage [ ] No Derivatives [ ]

From a datamodel point of view, this separates the name of the licence from the type of rights it provides, which will allow us to show in the discovery interface, for example, journals under BY licences whether they are C-BY or some home-grown version.

Take a look at the datamodel for OAG, to see how we would represent this: http://oag.cottagelabs.com/developers/api

[dommitchell]

RJ: I think we're talking about slightly different things, so need to clarify. The form says "do you supply article metadata", which I have taken to be a flag (true or false) as to whether the publisher makes stats available. But are you also saying that you want to capture those statistics?

DM: yes

what are we actually going to store in the database license value, eg 'cc-by' or 'equiv cc-by' what are we going to present to the end user in the discovery interface license type, eg 'cc-by' or 'equivalent to cc-by'

But yes, let's discuss this and your attribute idea on Monday.

[emanuil-tolev]

@dommitchell the prototype form has a (single) question

3) ISSN or eISSN *

Shall we make this 2 fields which ask for ISSN and eISSN separately (like in the current edit journal form), or do you only want publishers to give you one of the identifiers? Not much work (for them) to give you both so I'd suggest that.

We also have the optional validation (i.e. one must be present, but it could be either of them. Both missing fails validation though.).

[emanuil-tolev]

I'm under the impression that this form should be behind a login (i.e. you need an account to suggest a journal). I can't see where this has been recorded and it's not in this issue, so I'm recording it here - correct me if I'm wrong.

[richard-jones]

We don't know yet - there is currently no way for a publisher to request or create an account, so we need to determine that workflow. The old workflow saw the admins provide the account after the suggestion was made.

On 22 February 2014 17:14, Emanuil Tolev notifications@github.com wrote:

I'm under the impression that this form should be behind a login (i.e. you need an account to suggest a journal). I can't see where this has been recorded and it's not in this issue, so I'm recording it here - correct me if I'm wrong.

— Reply to this email directly or view it on GitHubhttps://github.com/DOAJ/doaj/issues/81#issuecomment-35808050 .

Richard Jones,

Founder, Cottage Labs t: @richard_d_jones, @cottagelabs w: http://cottagelabs.com

[dommitchell]

Correction: SemperTool created the accounts upon a suggestion being approved for the DOAJ

[dommitchell]

yes, please make it 2 fields so where possible, publishers can enter both.

On 21 February 2014 at 19:04 Emanuil Tolev notifications@github.com wrote:

@dommitchell https://github.com/dommitchell the prototype form has a (single) question

 > > 3) ISSN or eISSN *

Shall we make this 2 fields which ask for ISSN and eISSN separately (like in the current edit journal form), or do you only want publishers to give you one of the identifiers? Not much work (for them) to give you both so I'd suggest that. We also have the optional validation (i.e. one must be present, but it could be either of them. Both missing fails validation though.). — Reply to this email directly or view it on GitHub https://github.com/DOAJ/doaj/issues/81#issuecomment-35756213 .

[dommitchell]

I was thinking something along the lines of:

Choose your licence:

CC-BY
CC-BY-SA
CC-BY-NC
Other:

if "Other", does it require:

Attribution [ ] Share Alike [ ] No Commercial Usage [ ] No Derivatives [ ]

From a datamodel point of view, this separates the name of the licence from the type of rights it provides, which will allow us to show in the discovery interface, for example, journals under BY licences whether they are C-BY or some home-grown version.

Take a look at the datamodel for OAG, to see how we would represent this: http://oag.cottagelabs.com/developers/api

We are happy with this approach and would like it to be implemented.

[dommitchell]

Some notes on the new form. (We realise that what is there now is not final. These are only reminders for me and you.)

1. Please be sure to refer to the spreadsheet sent to you as well as the gForm, it contains information on it not represented in the form.
2. Both and header and footer text still needs to be added
3. Help text needs to be added
4. If yes, add URL questions need to be merged into question.

[emanuil-tolev]

CC-Zero not to be added.

@dommitchell : review list of licenses that will go on the form as choices

[nimphal]

Languages should be controlled vocabulary according to https://docs.google.com/spreadsheet/ccc?key=0AsW3CmDjW93kdGlRU21FNFZSRHhPaUk0ZmtUUHdGNnc&usp=drive_web#gid=0 so a list of all languages is needed.

[nimphal]

@emanuil-tolev Record where you took country-codes.json from in the README

[dommitchell]

@dommitchell : review list of licenses that will go on the form as choices

Please add CC-BY-NC-ND.

[dommitchell]

Please note the text of questions 19 and 24 has altered slghtly: https://docs.google.com/forms/d/1u0F-fVbEHySkTRXvJLkfd1s_ace9-nTzRYIWdJwSR68/viewform

[emanuil-tolev]

Please add CC-BY-NC-ND.

Done (not on test site yet, just a note that it's been done)

[emanuil-tolev]

"Does the journal embed machine-readable CC licensing information in its article metadata?" Please provide an example. For more information go to http://wiki.creativecommons.org/Marking_works

Is the google form missing a field for the example here? Do you want the user to input a URL into an additional text field below this question if they say yes?

(sorry @dommitchell for the duplicate email you'd get - I realised this should have been recorded in github, not in the least since I'll need to reread this whole issue when reviewing the form)

[emanuil-tolev]

I assume we still want a Captcha at the end of the form. Though it will be mighty annoying after such a long form :).

EDIT: just to clarify people won't lose any data if they fill in the wrong captcha, all they'd have to do is fill it in again and hit submit.

[dommitchell]

I assume we still want a Captcha at the end of the form. Though it will be mighty annoying after such a long form :).

Yes, please although if you think it is unnecessary, please say!

Is the google form missing a field for the example here? Do you want the user to input a URL into an additional text field below this question if they say yes?

Yes please! If yes, *Please provide an example:'

[emanuil-tolev]

Yes, please although if you think it is unnecessary, please say!

The big problem is that it's going to time out by the time they reach the bottom of the page. So if we use it, users must click refresh on it before entering it, otherwise it's guaranteed to be wrong.

I'm currently leaning towards a form without a captcha - I'm not sure that bots will be able to submit this form (or want to). The problem is generic bots which roam the internet, not anything targeted at the DOAJ specifically. If we do start getting junk submissions it's probably worth either:

a. using Google ReCAPTCHA with suitable instructions to refresh it (or even refresh it automatically for users, though I doubt this is easy to achieve) b. actually looking about for alternatives - believe me I've done this but there's just nothing that's less than a day of work at least to integrate and test properly. Captcha verification is definitely not a solved problem it would seem...

So for now I'd go for not having it and see if anything bad happens. If so, clean up the mess and put it on temporarily, then pick one of the options above.

[dommitchell]

Agreed , please proceed without it

[emanuil-tolev]

(I'll edit this comment rather than make new ones so come to the issue on github for the latest info. This is a list for myself to begin with, but we'll review it on Monday as part of releasing the new form.)

Some comments on the text around the form, taken from the prototype at https://docs.google.com/a/cottagelabs.com/forms/d/1u0F-fVbEHySkTRXvJLkfd1s_ace9-nTzRYIWdJwSR68/viewform :

All questions, except two, are mandatory.

Not sure if this is the case any more, just need to review.

A note about the qualifiers for the DOAJ Seal

Do we still want this section included at all?

(Question 17)

Er, the questions aren't numbered. They could be, but updating the form like that will be painful (have to change all references when adding / removing a question). That's not a great idea considering this form may be modified due to feedback very soon. Something which calculates the number of the question and inserts it dynamically into the text is better, as long as it does not take too long to develop.

What was the first calendar year in which a complete volume of the journal provided online Open Access content to the Full Text of all articles (Full Text may be provided as PDFs

Does this year have to be in the past? Currently max number allowed is current year + 1 (accommodate submissions around New Year's Eve I guess - your call though)

---

• Is the license URL absolutely required when the license picked is one of the CC options? What about the "Other" option? I'm guessing it's optional on "No"

[dommitchell]

All questions, except two, are mandatory.

Not sure if this is the case any more, just need to review.

DM: no longer needed. Please delete.

A note about the qualifiers for the DOAJ Seal

Do we still want this section included at all?

DM: yes.

(Question 17)

Er, the questions aren't numbered. They could be, but updating the form like that will be painful (have to change all references when adding / removing a question). That's not a great idea considering this form may be modified due to feedback very soon. Something which calculates the number of the question and inserts it dynamically into the text is better, as long as it does not take too long to develop.

DM: yes, I agree. If you could do that, it would be great! We definitely need the explanations there. If this can't be done before rollout, can we add targets in the text instead? e.g

1) have [an archival arrangement in place](hyperlinked to the question) with an external party. If 'None' is selected, the journal will not qualify for the Seal. etc

What was the first calendar year in which a complete volume of the journal provided online Open Access content to the Full Text of all articles (Full Text may be provided as PDFs

Does this year have to be in the past? Currently max number allowed is current year + 1 (accommodate submissions around New Year's Eve I guess - your call though)

DM: no, it can be the current year (for new journals or journals that are newly OA)

Is the license URL absolutely required when the license picked is one of the CC options? What about the "Other" option? I'm guessing it's optional on "No"

DM: no we decided to leave this optional.

[emanuil-tolev]

If this can't be done before rollout, can we add targets in the text instead? e.g

1) have an archival arrangement in place with an external party. If 'None' is selected, the journal will not qualify for the Seal. etc

Yeah, probably wouldn't try to do it today. And yes, targets will work, all the questions have one - good idea.

[emanuil-tolev]

DM: no, it can be the current year (for new journals or journals that are newly OA)

Do we allow next year at all then? Like now, should you be able to type in 2015?

[dommitchell]

Feedback from call with Lars: 1) Archiving, Article identifiers and Deposit policy - not urgent, but a suggestion would be to make the things clickable - that is the have a hyperlink below LOCKSS, CLOCKSS etc - especially the more exotic handle services and deposit policy services are not well known - this should by no means prolong the implementation of the new form

2) Question numbering: we will have to have the questions numbered. DOAJ Team will be communicating to publishers re problems with the form AND we will be translating the form in Help docs so question numbering facilitates that communication. The hyperlinked targets can stay!

3) 'Additionally I noticed that Article Identifiers should behave like the digital archiving policy, i.e. grey checkboxes out, "None" must be the only one selected.' DM: And move "None" to the top. 'Are there any other fields which need to behave like this actually? deposit policy? DM: yes actually. 'With which deposit policy directory does the journal have a registered deposit policy?' And move None to the top.

[emanuil-tolev]

Additionally, I've put up a new 2-column layout with the signature rounded orange boxes surrounding the sections: http://doaj.cottagelabs.com/application/new . Thoughts?

A couple of the headings (Basic journal information, Information about you) looked really strange when left-aligned, so I've aligned them (just them 2) with the where the text fields start.

If you resize your browser you'll see that the sections stack on top of each other in the correct order. (for smaller screens)

[dommitchell]

Yes, we like!

[emanuil-tolev]

Do we allow next year at all then? Like now, should you be able to type in 2015?

No. Just limit to current year.

[emanuil-tolev]

1) Archiving, Article identifiers and Deposit policy - not urgent, but a suggestion would be to make the things clickable - that is the have a hyperlink below LOCKSS, CLOCKSS etc - especially the more exotic handle services and deposit policy services are not well known - this should by no means prolong the implementation of the new form

OK, sure, though might take an hour or two (a. making the text of a checkbox a link; b. finding a maintainable way of defining those links in the code, we do have 57 questions and lots of checkboxes).

Probably best done with a list of items that you would like hyperlinked, and the relevant URL-s - you know far more than me which ones need a link.

Also, the links can be the standard browser blue or (similar to question numbering below) the DOAJ orange, if needed. I think it's also probably a good idea to open these links in a new tab :) considering you'd otherwise leave the huge form.

2) Question numbering: we will have to have the questions numbered. DOAJ Team will be communicating to publishers re problems with the form AND we will be translating the form in Help docs so question numbering facilitates that communication. The hyperlinked targets can stay!

Actually there was a compounded problem with the 2-column layout - the layout's very nice, but now you've got 2 questions next to each other on the screen. So you can't really tell what to look at when you follow some link to #deposit_policy since there's so much on the screen - it's gonna be at the top of your window but is it on the left or the right?

I thought we should really have a good solution to this, also because you raised the question numbers issue. There should be a GOOD way of linking to individual questions as that will result in much clearer feedback and less frustration. So

1. go here: http://doaj.cottagelabs.com/application/new
2. Hit Ctrl+F5
3. Click on one of the numbers next to a question.
4. Copy the URL (it should have changed to reflect the fact you clicked on it).
5. Open a new tab, paste the URL there (we're replicating the workflow of "I'll send the DOAJ a feedback email"). Hit Enter.

What should happen is: at 3. the page should scroll to the question (not snap to the question - scroll visibly so your eyes can follow). Then the question should get highlighted. at 5. the page will open at the position of the question, but it should get highlighted.

Additionally I've added the ability to refer to questions by their numbers. So http://doaj.cottagelabs.com/application/new#25 should work. This doesn't activate the fancy stuff (so it will just snap and not highlight), a. because it's a bit difficult to infer the name of a question from its number at the moment b. because presumably you know what you're doing if you use that link. There's only 2 ways to get it: read the source code of the page (I would do this to random websites) OR you just know about it.

It's obvious what you're looking for anyway, the numbers are on the screen...

3) 'Additionally I noticed that Article Identifiers should behave like the digital archiving policy, i.e. grey checkboxes out, "None" must be the only one selected.' DM: And move "None" to the top.

OK, done.

'Are there any other fields which need to behave like this actually? deposit policy? DM: yes actually. 'With which deposit policy directory does the journal have a registered deposit policy?' And move None to the top.

Done.

[emanuil-tolev]

As an update: there's a bit more work on the crosswalk as a result of the work on making questions referanceable + numbered. I will make it live tomorrow morning (10-11 UK time) unless there's any other feedback on the front-end.

Of course I'll send an email around to everybody when I do that.

[dommitchell]

Just to say thanks for the numbering solution! It's great!

[dommitchell]

Assigning to Emanuil

[emanuil-tolev]

Some updates on what has been developed between the comments above and now.

• The crosswalk! We have a big hunk of code that converts the 56-field form to an actual DOAJ application ready to be saved to the index.
• Tabbing from one field to the next with the keyboard was broken when question numbers were introduced (they are links, so the Tab key stopped at them). You can no longer navigate to those links with the keyboard.
• Error colouring was off. Errors themselves were black text, descriptions became red - quite nonsensical. Now descriptions remain black, error messages are bright red.
• Digital archiving question: "A national library" the extra text field had disappeared due to me and @dommitchell changing the wording of that option. Fixed and made it so future changes to the text of the option don't make the field go poof.
• Currency options have been deduplicated. Multiple countries use GBP and EUR, who would have known.
• Harmonised the values stored for License on the form with what we currently have. Otherwise we end up with "cc-by" and "CC by" and a big cleanup job again. We should clean up the values, but not now.
• Several URL-s in the descriptions of some fields were not actual links. Made them links, open in new tab.
• Making the scheme in URL-s optional was a very good idea, but adding http:// automatically was not. Specifically you end up with things like http://http:// if you leave URL fields empty (even optional ones!). So that was fixed.
• Several important examples were missing from descriptions / descriptions weren't in sync with Google form. That too was fixed.
• Keywords accepted more than 6 with Javascript off. Fixed.
• Cross-walk turned up several issues, like "do not record 'Other' as a valid value" for things like digital archiving policy and license. The contents of the Other text field are needed instead.
• The License URL field ( https://doaj.org/application/new#license_url-container ) was being hidden if the user answered "No". "No" means "not CC-like license", not "No license". They still need to be able to put in a URL if they say "No", though who knows why they're applying at all in this case.
• A very technical but quite serious detail due to the unexpected time it took - there were also instances of "True", "False" and "None" being recorded. As text, not logical booleans. Resulting in things like digital_archiving_policy of ["No policy in place"] - so hey, the journal has 1 policy! Which is obviously wrong, the journal has 0 policies (i.e. [ ]). These special values have all been translated, ignored and overall interpreted as needed.

Probably best to open new issues about new enhancements to the form or user feedback. I do note we have 2 suggestions on the live site as of now (~2 hours live) so it is at least working :).

[dommitchell]

Issue 132 opened: https://github.com/DOAJ/doaj/issues/132