Closed rybesh closed 9 years ago
Re: issue categorizations; yes these are fully customizable. Once we have a few more we can think about better ways to organize them.
As far as preserving original text, the intention for start/stop endpoint values is that the "label" property in the PeriodO JSON (see for example https://github.com/periodo/examples/blob/master/data.json#L58) should contain the original text parsed by the input interface. So we need to make sure that is what is happening, and perhaps make it clearer in the UI that the original text is being maintained.
As for the spatial coverage values: what we really need here are some "cataloging rules." How would you instruct a student to select spatial coverage values, given some text in a source? Once we can decide on these rules, we can come up with a way of representing the data that reflects them. Right now we have a kind of odd mix of mostly modern nations but a few things like "Roman Empire," and it's unclear why one or the other was chosen.
Thanks for moving this over, Ryan. It wasn't clear to me that the original labels were being maintained on either the front or back end, especially because the input interface didn't allow unrecognized formats (ca., etc.), but I gather Patrick is working on this.
I'm agreed about cataloguing rules for spatial coverage. The few ancient things are those for which spatial data (e.g. polygons or points) is available for representation (I think there's a URI for "Roman Empire" with spatial extent that we'd identified, though obviously that's also time-dependent, and I included some ancient sites that have URIs in Pleiades). I think maintaining the ability to display things on a map is important, so my criterion for choice here would be "the online data source that is the most likely to have a spatial extent that matches the concept in the mind of the source making the definition", whether that's a set of modern countries or ancient imperial boundaries". The latter only works when there's a defined political entity, so "Mesopotamia" is more likely to be rendered by the former.
The initial dataset -- with textbooks -- is going to be the trickiest, since these are the broadest. Most of our thesauri are already applied to particular modern countries, and more specialized archaeological or art-historical literature is likely to be more specific about space as well. It's the generalizing works that are hardest to pin down. Maybe I should come up with an equivalency chart (e.g. "Northern Mesopotamia = Turkey, Syria, Iraq"; "the Levant = Syria, Jordan, Lebanon, Israel, Gaza and the West Bank"). What are your thoughts on this?
I think an equivalency chart is an excellent idea. As for "the online data source that is the most likely to have a spatial extent that matches the concept in the mind of the source making the definition," it might be worth making a ranked list of such sources, i.e. first check Pleiades, then DBpedia, then Geonames, etc. The order might vary by type of spatial entity. But I think the more we can try to translate your/Eric's feeling for what to do into explicit heuristics, the better.
Using the new client v.0.3.1, the data entry for an individual periodization currently lacks the following fields: alt-label, label_en, spatial_coverage_label, and notes_original. And source_locator, which seems particularly important. See these two screenshots: and
This appears to be the case whether or not I'm signed in to my ORCID iD. I guess I'm not sure if leaving these fields out is intentional and why if so.
Sounds like possibly Sarah is using a cached older version of the client?
On Mon, Mar 30, 2015 at 11:00 AM, Sarah B notifications@github.com wrote:
Using the new client v.0.3.1, the data entry for an individual periodization currently lacks the following fields: alt-label, label_en, spatial_coverage_label, and notes_original. And source_locator, which seems particularly important. See these two screenshots: [image: 20150330_collection_periodization] https://cloud.githubusercontent.com/assets/10801404/6898887/6a84091a-d6c2-11e4-8924-8ba7f42845fc.jpg and [image: 20150330_altlabel] https://cloud.githubusercontent.com/assets/10801404/6898896/7aea564c-d6c2-11e4-9822-0ddef53a7634.jpg
This appears to be the case whether or not I'm signed in to my ORCID iD. I guess I'm not sure if leaving these fields out is intentional and why if so.
— Reply to this email directly or view it on GitHub https://github.com/periodo/periodo-client/issues/2#issuecomment-87714324 .
Should I have her empty the cache and reload? We're still using test.perio.do, right?
Did the idb just go offline?
Sorry, it appears just to be Chrome, which I've broken somehow.
Also, yes, it does look like Sarah's using a cached older version. If she clears the cache, will she lose all the entries she's added to her local copy?
I'm pretty sure that a Ctrl-R or Command-R will refresh the JavaScript but not clear the local storage, BUT let's wait to hear from Patrick before she does that.
On Mon, Mar 30, 2015 at 1:43 PM, atomrab notifications@github.com wrote:
Also, yes, it does look like Sarah's using a cached older version. If she clears the cache, will she lose all the entries she's added to her local copy?
— Reply to this email directly or view it on GitHub https://github.com/periodo/periodo-client/issues/2#issuecomment-87767444 .
Hi ... I was using the older version via this URL: http://perio.do/periodo-client/# . Whether it was cached or not I am not sure, but the presentation there is clearly the old version and I'll drop that for future work. Now I'll use the new v0.3.1 to add period collections & periods via this URL. https://test.perio.do/#p/periodo-client/ (that backend; web doesn't seem to allow add/editing). Hopefully that sounds OK. (All on Firefox browser).
On Mon, Mar 30, 2015 at 12:43 PM, atomrab notifications@github.com wrote:
Also, yes, it does look like Sarah's using a cached older version. If she clears the cache, will she lose all the entries she's added to her local copy?
— Reply to this email directly or view it on GitHub https://github.com/periodo/periodo-client/issues/2#issuecomment-87767444 .
Working in https://test.perio.do/#p/periodo-client/ , I'm able to add a periodization within the Twist period collection. Then once it displays in my local backend, the Edit button disappears, till I switch to the JSON-LD view, switch back to List view, and the Edit button reappears. But it's nonfunctional. So it looks like one can only add new period collections while working in a local backend, which I'll do next.
[Firefox] OK, I was able to add two periodizations in a NEW period collection, then the "Add period" button disappears after clicking it to add a third one. I'm still in the local backend but can no longer add any more periodizations. Refreshing brings me back to the local period collection list but then, still unable to add to it. (If it adds anything, clicking "Sign in" at the upper right is also nonfunctional). Please advise.
I see what's going on. Some changes I made broke that editing functionality. I will fix it this afternoon.
@sarahb, the problem of the add button is fixed.
The original start/stop text issue remains outstanding
Hmm ... the "add period" button still has a bug: this time I was able to add 5 periodizations within a new collection before the "add period" button disappears. See screenshot:
After five (v0.6.5; using v0.3.1 before I got through two), the page completely freezes up, and all 3 Export options are nonfunctional too. View as JSON and View as RDF don't work, View Visualization blanks the screen / View List returns it back. Only thing that works is to "switch" the backend, choose the same local again, then I'm returned to the period collection page and can continue adding more periodizations to it...
UPDATE: 5-periodization limit? When the "add period" button reappears, I click it and the page freezes again with no entry options for more periodizations. I wanted to get through adding 10-20 periodizations to this April22 patch, but it doesn't allow me to add more after five, so I'm going to submit a patch of five.
Hi Sarah,
Thanks. The five period limit must be significant. I'm not going to be able to work on anything until Friday afternoon, but I will tackle this then.
The start/stop display issue is now fixed.
I don't know what the five-period limit was all about, but there has been a complete overhaul since then. That problem does not exist anymore.
via @atomrab: I wanted to move our discussions into the issue tracker, since I've found that as productive, if not more so, in other projects than long email threads.
I don't know if one can add categories for labels for issues, but I'd like to have something to set apart actual bugs, enhancements, and questions from larger theoretical issues. Does anyone have a suggestion for a tag we can put in a title for bigger-picture issues like this one? Or is this just really an enhancement?
Now on to the issue.
I wanted to submit this issue to check in with you and Ryan about the display of the periods entered in the client you're building. One of the principles I assume we want to maintain is the preservation of the original terms in which a given period was described (so even if we're representing it as ISO dates, we keep the original statement that it was "early 5th century BCE-25 CE" etc.). Right now it looks like BCE, for example, gets translated to BC, and ISO-style dates (e.g. -700) aren't recognized by the parser.
Obviously we'll need to normalize the dates on the back end and in the representation, but I'd find it helpful if the edit/browse interface retained the original label for display, either alone or in combination with the normalized ISO dates. I think it makes most sense to be explicit about any transformations that are happening between the temporal coordinates as expressed by the source and the information we're collecting.
Right now I think it's impossible not to mask the spatial transformations, though -- so when someone says that a period applies to "Mesopotamia", we're stuck making some deductions about the modern nations to which that applies. It will be interesting to see if we can get anywhere with ancient gazetteers or data sources -- Pleiades or maps of the extent of the Roman Empire, for example. But here too I'd be in favor of being as explicit about deductions as possible.