Should we make ODM WP contents accessible on the search in country pages?

[acorbi]

Currently, users can head to the search page on both ODM and country sites in order to find different types of content they might be looking for. The difference between the search on ODM and the search on a country site, say ODC, is that the search in the country sites is pre-configured so only contents related to Cambodia are exposed.

In case of CKAN content (datasets, library publications, laws records and agreements) there is no problem because editors can specify the country or countries each record relates to, while filling out the metadata.

This is not the case for WP content where the definition of which country relates to the content is being made automatically. Contents on ODC WP will be automatically tagged with Cambodia, which makes sense. Contents on ODV will be automatically tagged with Vietnam...

The problem arises with content stored on ODM WP that might relate to Vietnam and that should be made discoverable on the ODV search. An example would be a certain news article posted on ODM that refers to some event on one of the local countries, like https://opendevelopmentmekong.net/news/vietnams-social-media-shaping-new-environmentalism/. This record will only be exposed as part of the search results when the user triggers the corresponding search on ODM, but not on ODV.

In order to solve this issue, following options can be considered:

Making all ODM WP content accessible on the search page of country sites

The search on the country sites can be tweaked in order to show WP content stored on ODM, besides the content stored on itself. This would make possible that, for instance, a certain news article posted on ODM which relates to vietnam (see example above) appears on ODV's search if the user specify the matching criteria.

The main PRO of this solution is that it does not involve major development or editorial efforts. The CONs are that, depending on the search criteria specified by the user, certain content might appear on the search results, even if it is not related to the country site the user is visiting.In the case of the example above, a user could enter Vietnam social media on ODC's search and the result would appear.

Enabling ODM editors to specify which country the WP contents relates to

With a little amount of development costs involved, the WP Dashboard can be extended so ODM editors could specify the country or countries a certain piece of content is related to. Similar to the data entry process in CKAN, WP editors would be able to specify one or more countries for each content piece they work on.

Obviously this solution involves manually adding this information to those ODM posts (news articles, topics, etc..) that might be also exposed on the corresponding country sites. Since ODM has a considerable amount of contents already, this operation could be then planned only for content to be added in the future or specific contents which could be edited individually.

NOTE: Until some time ago, editors were using WP Tags to indicate which country or countries the news articles were associated with. One task of this proposed solution would be to automatically assign this information based on the tags used on previously uploaded content.

@OpenDevelopmentMekong/editorial @OpenDevelopmentMekong/tech @OpenDevelopmentMekong/partnerships @OpenDevelopmentMekong/odc @OpenDevelopmentMekong/odmm @OpenDevelopmentMekong/data Please provide your thoughts

[acorbi]

Quoting @jpeizer1 :

Perhaps the solution would be to clearly identify ODM returned data in the search and provide a filter in the country search screens to filter ODM data out. That way you can automate the process without having editors tag each item and allow the users to filter out the ODM returned results if they don’t wish to see them when doing searches in country. The filter would be something like: “Exclude regional results from Mekong Site”

[gimmemochi]

@acorbi Correct me if I'm wrong: WP content are: news, public announcements, and briefing pages (topic and landing pages).

I agree that relevant (to country site) content on ODM WP should be discoverable on the country site, and this is important in the cases where country site has few content.

However, selecting news as an example is tricky -- OD Mekong site is not supposed to aggregate country-level news. Country-level news should be aggregated and published on the country sites. (cc: @DBishton).

If we choose your proposed solution one "Making all ODM WP content accessible on the search page of country sites," we should prioritize country site content (WP and CKAN content). In cases where there's no relevant country site, then the search should return relevant content (WP and CKAN) from the Mekong site.

[ChandaraKhun]

@acorbi

Making all ODM WP content accessible on the search page of country sites Problems may appear in the following situations: -The cross-boundaries issues such as dams, water governance, illegal wildlife/log traffics. It may happen that each OD platform may have redundant content. -Behavior (s) of OD Mekong/OD country on publishing news & ODM on importing news from OD Cambodia

Should we make ODM WP contents accessible on the search in country pages? Beside the above mentioned issues, it is even harder for OD country/ODC to ensure its compliance to the editorial standards of bi-lingual content.

ODC Wordpress content: News, Announcements, Topics, Site updates, Profile and Download Maps.

[acorbi]

Thanks @gimmemochi and @ChandaraKhun for your feedback.

ODC Wordpress content: News, Announcements, Topics, Site updates, Profile and Download Maps.

Correct

However, selecting news as an example is tricky -- OD Mekong site is not supposed to aggregate country-level news. Country-level news should be aggregated and published on the country sites. (cc: @DBishton).

This might need some clarification @DBishton

If we choose your proposed solution one "Making all ODM WP content accessible on the search page of country sites," we should prioritize country site content (WP and CKAN content). In cases where there's no relevant country site, then the search should return relevant content (WP and CKAN) from the Mekong site.

mmm not sure if this is implementable, need to look at whether this is possible

The cross-boundaries issues such as dams, water governance, illegal wildlife/log traffics. It may happen that each OD platform may have redundant content.

@ChandaraKhun This is actually not a big risk since in case of duplication, there would be only the content coming from ODM and the content coming from the country site. Content stored on other country sites would not be exposed on search in country pages, only content from ODM and the current's country site.

Behavior (s) of OD Mekong/OD country on publishing news & ODM on importing news from OD Cambodia

Again, to clarify with @DBishton

Beside the above mentioned issues, it is even harder for OD country/ODC to ensure its compliance to the editorial standards of bi-lingual content.

This is also a risk, since ODM content would always appear in english only.

[acorbi]

@ChandaraKhun @Huyeng @gimmemochi @prustar @DBishton @thetaung2 @DragonVirus

After having a related discussion during SCM, we agreed on piloting a possible solution for this. I would like to kindly request all of you to provide input since these are decisions that affect all sites.

Currently on PP, I have implemented a mechanism to expose ODM content to the search of country pages. The mechanism basically makes it possible for the search to list results from ODM on country sites. Following details:

• This only involves exposing ODM content to the search on country sites. If users are searching on ODC, no content from other country sites (ODMm, ODV) will be queried.

• The search automatically adds the name of the country on the search query. That means that if a user searches for Land on ODC, the search engine will append the term cambodia to the search query. This should increase the chance that the ODM results presented are actually related to the country site to avoid that unrelated content is presented.

• The same applies for the suggestions and spell-checking mechanisms

Please review the current implementation and do some tests, your feedback on this is necessary in order to find a solution that works for everybody.

[Huyeng]

I would like to know the reasons behind of of the exposing the ODM content to the search page of the country, it would be great if any one can provide the explanation clearly on that. What are the benefits of the exposing for ODC site? What I have known, the exposing the datasets from CKAN through the geographical area into the country search page. Even thought, this has already been implemented, ODC is still discussing and studying about the pro and con, then make decision later for this.

Regarding to the exposing the ODM content to the search of the country site, I can see that it may good for other country sites that have less/no contents, but not ODC. As ODC has a lot of contents and datasets that our team published daily, exposing the ODM contents into the ODC will increase the number of contents/records for searching, as a result, searching would not be found as expected due to of many posts/records. You may already aware that the users frequently view the results on the first page or two or three of the first page.

Moreover, ODM has used the script to import ODC's news into the site, if exposing, it will show duplicately. In addition, the news and announcement of ODC and ODM has input and present in different format. An example, ODC's news no thumbnail, but ODM has.

The bi-lingual issue will be existed on the site, and it also an issue that is very concerned for us since ODC tried to claim to the Donors and our users that the contents on ODC are available in both languages. Besides the contents are not presented in local language, it will show a large different number of searching results between English and Khmer site which was an issue that the testing users raised.

Furthermore, related to topic type, if users search any of topic page "Energy", it will show up twice, the users will questions about the results because they don't know that the pages are different (ODC & ODM).

As raising the challenges above, for now, ODC don't want to expose contents of WP ODM into the search page. If the problem can be solved, then ODC will consider it again.

Interconnection of the contents across the boundaries is a good idea, but integrating the contents of an organization with other organizations has produced some questions as below:

• What are different between those websites?
• If each site has almost the same contents on searching, why don't we have a website?

[acorbi]

@Huyeng sorry for my late reply to your comment. I just saw it.

I would like to know the reasons behind of of the exposing the ODM content to the search page of the country, it would be great if any one can provide the explanation clearly on that. What are the benefits of the exposing for ODC site?

Very easy. If ODM has content related to Cambodia, It is logical to try to expose it also on ODC's search, as long as that content is not already on ODC. The origin of this discussion was that piece of content related to Vietnam which David Hindley asked testers to look for and which was actually on ODM and not ODV. The same could potentially happen with Cambodia-related content.

Now I see you are presenting a series of arguments that might be a reason for not implementing this on ODC. Let's look at them in detail.

Moreover, ODM has used the script to import ODC's news into the site, if exposing, it will show duplicately

@Huyeng @DBishton can you please elaborate on this? When did that import happen? How many records might be duplicated? Of which content type?

In addition, the news and announcement of ODC and ODM has input and present in different format. An example, ODC's news no thumbnail, but ODM has.

This is not a strong argument, is it? The list layout that we currently re-use across the platform renders well with or without thumbnail... Also, you can configure it to not show the thumbnail if you want it.

The bi-lingual issue will be existed on the site, and it also an issue that is very concerned for us since ODC tried to claim to the Donors and our users that the contents on ODC are available in both languages.

I agree this is an argument to consider.

Furthermore, related to topic type, if users search any of topic page "Energy", it will show up twice, the users will questions about the results because they don't know that the pages are different (ODC & ODM).

I agree this might confuse users

Interconnection of the contents across the boundaries is a good idea, but integrating the contents of an organization with other organizations has produced some questions as below: What are different between those websites?

This is my rationale: ODM is the regional site. It contains information which is relevant to the region (including cambodia). If there is content on ODM that is not available on ODC, then it is worth to exposing it on ODM as well.

If each site has almost the same contents on searching, why don't we have a website?

The thing is that ODM should not have the same contents as ODC. it might have similar or related content. But this is rather an editorial @DBishton or data @prustar related issue.

Since I understand your position and observe that there are arguments against exposing ODM results on a certain country site's search (in this case ODC), I am happy to implement following configuration mechanism described on https://github.com/OpenDevelopmentMekong/wp-odm_solr/issues/151

[Huyeng]

Thanks for replying.
Well, I just would like to explain further more on the issue of "Searching the content related to Vietnam which David Hindley asked testers to look for and which was actually on ODM and not ODV" to be clearer, event thought, I didn't join the testing, I also was there in the during feedback providing. Regarding to the issue, a tester really provided the comment (in Khmer) like that because they can't find the dataset related to the Vietnam on the ODV search page which was presenting on the old template. With the old design of the template, the result of CKAN just listed only title in the right column which no one considers as the results of searching including ODC team members.

The test tried to find "Vietnam Forest Law 2004", and it was found it in ODM, but not ODV.
The "Vietnam Forest Law 2004" is the dataset of ODV, so it must be available on in ODV website. That's what the users expected.

However, his comments has reflected to us to consider if it is a good idea or not to include the regional contents related to Cambodia in ODC site since there are some challenges that will be facing.

From my understanding, it wasn't expected that all the regional contents should be found in country searching page, but maybe only the regional contents related to the country that is visiting should be.

@Huyeng @DBishton can you please elaborate on this? When did that import happen? How many records might be duplicated? Of which content type?

I just would like to recall the MultiI-mporter script that has been implemented on ODM site since earlier. The script has been used import the news from ODC site as well as other vai RSS feed.

• The feeds will be updated every 30 mins
• Maximum number of items to import for each feed is 40
• The 10 latest of posts will be posted
• The posts status are drat (they will be published by any one from ODM) See this: https://opendevelopmentmekong.net/wp-admin/admin.php?page=wprssmi_options5

[prustar]

Hi All, just joining in on this discussion and I have to agree with Huy Eng that this issue raised is not related to WP content it was an ERRONEOUS dataset that was uploaded incorrectly during the time when datasets where being transitioned over to CKAN. I mentioned this in the search feedback that David provided. It is not a good example to provide for the context of this issue that we are talking about.

In regards to the news on ODM I agree with @gimmemochi that ODM should not be aggregating news from the country level sites. This defeats the purpose of having a regional site that offers a regional perspective that adds value to the national sites. If this import script is still in function then I strongly advocate that it be discontinued @DBishton - Please clarify.

I also have to agree with @Huyeng as well that the actual benefit of cross populating WP content between national and regional sites could be very redundant for all of the reasons she gave. I propose that perhaps we need to be creative about offering options to our users to explore while they are on our network of platforms. For instance I have previously raised the option of having a pop up box that provides some guidance on searching for content. For example: "not what you are looking for? try going to the regional Mekong site" or inversely "try visiting the national country site". I recognize that this would need some development to intuitively write the script but makes it easier to keep WP thematically different, we want people to visit the country sites and we want people to visit the regional site to get a national vs regional perspectives.