OAI does not export ISSN field

[dankomed]

Dear Team,

I have registered at OCLC, and have created repository at WorldCat Digital Collection Gateway.
For this I had to insert the OAI base url and then synchronize the metadata with my ojs website.
Unfortunately, I see no exported ISSN data. My question is whether this is the normal and expected behavior of the OJS installation, or there is some problem that I need to fix in my ojs installation. For example, I have inserted ISSN only as electronic, not print. However, I tried to put in both electronic and print ISSN fields the same number, then synchronize with WorldCat without success.
Any ideas how to sovle this problem? The WorldCat support team told me that this is a problem of my OAI settings, not their website.

[asmecher]

Hi dankomed,

There is already a Bugzilla entry for this feature request; see http://pkp.sfu.ca/bugzilla/show_bug.cgi?id=7392 (and consider CCing yourself on the bug entry so you're notified of status changes).

Regards,
Alec Smecher
Public Knowledge Project Team

[dankomed]

Dear Alec,

I registered with Bugzilla and added myself as follower of the reported bug. I was not able to understand however, whether the bug is already solved in the new OJS version (I am using 2.3.7) so that I actually need to upgrade, or the bug is in the waiting list to be solved, and I have to wait.

I am not a professional developer of software, rather I am an experimental and theoretical scientist, but nevertheless I am testing various free software programs in need, and do have ideas on customizing or writing myself new code, based on modifying portions of already existing code. Here I would like to share some observations on the functionality of the OCLC digital collections website, and I propose an idea of what can be implemented.

In OCLC they provide option to set "constant" fields for the whole repository. For example, I can set language for articles to be "English", and then "map" this customized field to specific field in the OAI data. With possible typos - OCLC gives you option to create field which is something like dc: identifier: 2 = xx or dc: language:2 = xxx. The OAI exported fields are usually numbered as :1, so these new OCLC generated fields have larger number e.g. 2. Once I create such new field say dc: identifier: 2 = 1314-7374 I can next map it onto the ISSN field into OCLC. It is displayed perfectly OK in WorldCat database, but the WorldCat search function for ISSN does NOT retrieve the entries by the ISSN. The problem is in that the system does not recognize that the entry is for ISSN i.e. "dc: identifier: " should be replaced by code specifying that the field is for issn.
I was wondering whether OAI can be configured easily to export at least empty ISSN field, then possibly OCLC site will automatically provide option for overwriting this field with "constant" one. At present it seems, OCLC provides such overwrite with "constant" field only for already existing OAI data fields.

Sorry for the layman description. I have no idea about the proper terminology that software developers use for any of the above.

[asmecher]

Hi dankomed,

I've just committed some code to that bug entry; that feature is not in any released versions of OJS but future releases will include it. A patch should pop up there in a sec if you'd like to try it.

Note that dc:identifier probably isn't a good fit for the ISSN, since dc:identifier should unambiguously identify the article, and ISSNs only identify the journal. The dc:source element is a better match (and that's where the patch puts it).

Regards,
Alec Smecher
Public Knowledge Project Team

[dankomed]

Dear Alec,

I wanted to implement your code but unfortunately my OJS version is 2.3.7 and does not have the dc11 folder inside plugins/metadata. I suppose copying the whole folder dc11 from version 2.4.1 to 2.3.7 will not work without additional change. If you think it would be easy to write down the necessary steps to make the dc11 folder actually work under ojs 2.3.7 I would like to do it.

The upgrading from OJS 2.3.7 to 2.4.1 failed two times. I am using GoDaddy server, so unfortunately I cannot implement the patch using the command mode. I tried to use the steps with the complete ojs 2.4.1 - update the config.inc.php file, copy the /cache and /public folders from the old ojs, turn off the installation and go for upgrade. In all cases, the mysql database is corrupted, and I have to restore it. In one of the trials, the upgrade procedure seemed to be OK, but then the articles in the first published issue could be found - rather error messages appeared.

I think the upgrade procedure could be perfected for users who do not have access to command line, but I suspect that this is not something of high priority for the OJS Team. That is why, I would like to know if there is a way to upgrade only the OAI plugin, possibly by copying the dc11 folder from OJS 2.4.1, and then configure the 2.3.7 system to actually use it.

Thanks for your help.

[dankomed]

Dear Alec,

I tried connecting to GoDaddy server by SSH, but they have disabled the patch command. After login to my server PuTTy gives the error: "bash: patch: command not found" and in forums GoDaddy admin replies: As indicated by the error message, the patch command is not supported within SSH on our Shared Hosting plans. Only within our Virtual Dedicated Server or Dedicated Server accounts would you be able to use this command via SSH.

So, I am unable to patch the ojs, and cannot afford paying for dedicated server. I would be happy to know how can I update only the OAI manually from 2.3.7 to 2.4.1 so that I can export ISSN. Or is there a simpler patch to export ISSN that works specifically for ojs 2.3.7? Thanks.

[asmecher]

Hi dankomed,

Two options for working around a disabled or unavailable patch tool:

• Download the affected files to your local system and apply the patches there (depending on your system, you may already have GNU patch on it [e.g. MacOSX or Linux], or may need to download it), or
• Make the changes manually using a text editor.

Either will require you to have a passing familiarity with the patch file format. Lines beginning with +++ and --- describe what file is being changed, lines beginning with + describe additions to that file, lines beginning with - describe removals, and lines beginning with @@ tell you roughly where in the file you should be looking. There is lots of information about this online.

Regards,
Alec Smecher
Public Knowledge Project Team