Skip to end of metadata
Go to start of metadata

Summary

We went through the first iteration of ROBox with Marco and Pique to gain some feedback. The comments/discussions raised were:

  • Confusion between Dropbox folder vs Research Objects Location vs Research Object folder.
  • UI needs to be much clearer; terminology is cryptic and non-intuitive at first glance.
  • The information on the Dropbox connection should be separated out from the information on Research Objects. The latter is much more important to users.
  • Suggestion is to have a "My Research Objects" section.
  • The common case is for users to have a single "ROs Location", not multiple. So the ability to add multiple should be hidden away as a power feature.
  • Users liked the folder sharing aspect of Dropbox+ROBox as it allows for easy collaboration and also to ease people into using Research Objects.
  • Missing: ability to see and edit metadata about the ROs and their contents.
  • Missing: ability to define links to external resources.
  • Reusing other people's ROs; Forking and cloning ROs.
  • Removing ROs and whether ROs should actually ever be fully deleted.
  • Issues with sharing data / privacy and when to "put it in" an RO and sync with the backend.
  • Have to take into account organisational policies.
  • User request: ROs should contain info on the creators (and in general the metadata capabilities must be extensible).
  • Information adding on submission and in environments like Taverna.

Chat Log

[13:01:56] *** Jiten Bhagat added Marco Roos ***

[13:02:45] *** Jiten Bhagat added Stian Soiland-Reyes ***

[13:02:48] Jiten Bhagat: Hello

[13:03:19] Pique: hello

[13:04:25] Jiten Bhagat: Thanks for joining Pique

[13:04:29] Jiten Bhagat: we'll wait a few mins for Marco

[13:04:40] Pique: ok

[13:05:18] Jiten Bhagat: in the meantime, let me share a folder with you on Dropbox, to allow you to put your ROs in them

[13:05:31] Jiten Bhagat: any preferences for a folder name? is "PiqueROs" okay?

[13:05:43] Pique: hahaha

[13:05:49] Pique: put AstroROs

[13:05:54] Pique: think it's better

[13:06:14] Jiten Bhagat: ok :)

[13:06:33] Jiten Bhagat: what email address should i share it with?

[13:06:45] Pique: xx@xxxxx

[13:07:45] Jiten Bhagat: ace, thanks

[13:07:59] Pique: each RO will have its own folder inside AstroROs ?

[13:08:08] Stian Soiland-Reyes: yes

[13:08:10] Jiten Bhagat: yeap

[13:08:13] Stian Soiland-Reyes: each folder you make will become a new RO

[13:08:18] Pique: ok

[13:08:26] Stian Soiland-Reyes: well, the direct children - an RO can have folders inside it

[13:08:29] Marco Roos: Kostas and Eleni are joining me.

[13:08:33] Jiten Bhagat: great!

[13:08:37] Marco Roos: Need a minute to set up properly

[13:08:54] Jiten Bhagat: Marco - what should i call your ROs folder, and who should i share it with?

[13:08:55] Jiten Bhagat: sure

[13:11:58] Jiten Bhagat: (once we get set up, Stian and I can give a quick overview of what we have at the moment)

[13:12:08] Marco Roos: Are we doing a skype call, or telcon, or just chat?

[13:12:17] Jiten Bhagat: just text chat, if that is okay

[13:13:02] Marco Roos: Good, no need to set all this up then (blush) ;)

[13:13:09] Jiten Bhagat: :)

[13:13:47] Marco Roos: Eleni and Kostas are coming in

[13:14:03] Marco Roos: My laptop is on screen

[13:14:07] Jiten Bhagat: great

[13:14:10] Marco Roos: so you can share stuff with me.

[13:14:32] Jiten Bhagat: @Marco - is "BioROs" okay for your research objects folder? (the one that will contain ROs in it)

[13:14:55] Marco Roos: Sure, so I find that somewhere in my DropBox?

[13:15:24] Jiten Bhagat: I'll share that with you (and anyone else you'd like)

[13:16:03] Jiten Bhagat: @Marco - what email address should i share the folder with?

[13:16:28] Pique: ok, in my case you can also share AstroROs with xxx@xxxx and xxx@xxxxxx

[13:16:30] Marco Roos: xx@xxxxxx

[13:16:37] Jiten Bhagat: thanks

[13:16:38] Marco Roos: (they all link to one address)

[13:18:05] Jiten Bhagat: @Pique - if it's okay with you, I'll share with a broader audience at a later stage; the current set up of sharing folders with you is just temporary (because we have to use a sandbox account only for now). In the future you'll be able register your Dropbox account directly

[13:18:15] Marco Roos: I have accepted; see an empty folder

[13:18:30] Pique: ok, no worries

[13:18:39] Stian Soiland-Reyes: @Marco - good point - should there have  been a README?

[13:19:15] Stian Soiland-Reyes: quick, Jits, add it!

[13:19:17] Pique: I see a .dropbox file with a number inside

[13:19:18] Jiten Bhagat: http://robox.wf4ever-project.org/ is the "Dropbox connector"

[13:19:19] Marco Roos: uhm, I guess so; we don't know the purpose just jet.

[13:19:34] Jiten Bhagat: at the moment, i havent connected up these shared folders with Robox, so no sync will happen

[13:20:05] Jiten Bhagat: and Robox only works with the sandbox wf4ever dropbox account (for now)

[13:20:16] Jiten Bhagat: (apologies if this is all confusing, we'll do our best to clear this up)

[13:20:35 | Edited 13:20:45] Jiten Bhagat: the aim today is to get you creating some research objects and take it from there

[13:21:19] Marco Roos: ah, ok, we'll try to get in now.

[13:21:41] Jiten Bhagat: what we have at the moment is very basic -  it sync ROs that you create within a designated "ROs location" (ie: the folders i have shared with you)

[13:21:46] Jiten Bhagat: with the dLibra backend

[13:21:51 | Edited 13:21:52] Jiten Bhagat: and creates a manifest RDF file

[13:22:03] Jiten Bhagat: I'll now connect up your RO locations with the robox app

[13:22:28] Marco Roos: what does that mean?

[13:23:16] Jiten Bhagat: (sorry, one sec)

[13:23:23] Jiten Bhagat: I thing the dlibra backend may be down :()

[13:23:29] Marco Roos: hurry, hurry! ;)

[13:24:02] Pique: @Jits please be clear in which are the things users should do which are the things that only you as developer are doing for this demo

[13:24:35] Jiten Bhagat: ok, will do

[13:25:04 | Edited 13:25:16] Marco Roos: @jits are you sure you don't want to talk?

[13:25:04] Jiten Bhagat: one sec, seems like that robox app can't connect to the backend, i need to set up some onfiguration

[13:25:21] Jiten Bhagat: not unless you want to hear my cursing at the software Marco ;)

[13:25:56] Marco Roos: majority vote: if you can do that in Malawian, then yes!

[13:26:46] Pique: it's the first thing we learn from a foreing language ;)

[13:26:59] Marco Roos: Kostas wonders what the interface for the user will be: Robox or DropBox?

[13:27:15] Marco Roos: but maybe we will see through this chat...

[13:27:58] Marco Roos: I missed your message about dLibra backend; you working on that then?

[13:28:23] Jiten Bhagat: hehehe

[13:28:27] Stian Soiland-Reyes: as always with live demos

[13:28:43] Jiten Bhagat: @Marco - i guess it would be both, but we won't know until we've tried that out

[13:28:59] Jiten Bhagat: yeah, sorry about this, looks like something has changed in the dlibra backend

[13:29:05] Jiten Bhagat: which has caused this to break :(

[13:29:29] Jiten Bhagat: in the meantime, I'll do a better job of explaining what we have :) ...

[13:30:07] Jiten Bhagat: the aim is to allow users to have dropbox folder(s) in which they deposit and work with research objects (as a collection of files and sub-folders

[13:30:42] Jiten Bhagat: then, users can connect their dropbox account with dlibra using an intermediate thing - Robox

[13:32:02] Pique: do they need RObox every time they want to syncro ?

[13:32:03] Jiten Bhagat: in the "real" situation:

1. users would go to the robox site

2. sign in with their personal/work dropbox account

3. select a location that has ROs

4. then let robox synchronise their current ROs and future ROs/updates with the dlibra backend

[13:32:21] Jiten Bhagat: the dlibra backend will then create and manage an RO manifest

[13:32:32] Jiten Bhagat: @pique - no, that will happen in the background

[13:32:37] Jiten Bhagat: and is a one time setup

[13:32:43] Pique: great !

[13:32:48] Jiten Bhagat: but it's possibe that Robox can also be a user interface to these ROs

[13:33:28] Jiten Bhagat: esp, to do things like:

- edit the metadata about the RO (title, desc, tags, etc)

- add and work with remote links

- get a "dashboard" over your ROs

[13:34:06] Marco Roos: will I, as a user, first designate DropBox folders as RO's, or will Robox create DropBox folders?

[13:34:36] Stian Soiland-Reyes: @Jiten - or perhaps such a user interface is separate from Robox - after all your ROs could come from many places

[13:34:42] Stian Soiland-Reyes: kind of like myExperiment for ROs

[13:35:06] Jiten Bhagat: @Marco - in robox you register "RO locations", these will be the "containers" of ROs and can already exist (if they dont exist then Robox will auto create them in your dropbox), then any new folders within them will be treated as individual ROs

[13:35:20] Jiten Bhagat: @Stian - yeap, there could be multiple interfaces

[13:35:57] Marco Roos: will I be able to screw things up on either side (robox for dropbox, dropbox for robox)?

[13:36:05] Jiten Bhagat: currently, we are in "sandbox" mode, which means only one dropbox account is allowed to be connected (for technical reasons), hence why i have shared these folders with you

[13:36:06] Stian Soiland-Reyes: you will have the power to screw up on either side, I guess

[13:37:05] Jiten Bhagat: if you would like to log in to Robox right now, you can do so by first logging out of the dropbox website, and then using these credentials when dropbox asks you to log in:

Username: xxx@xxxxx

Password: xxxxx

[13:37:06] Stian Soiland-Reyes: robox won't touch any other folder than the one you register. Every direct folder in there (if it exists) will be registered as a new RO. Any direct files will be ignored.

[13:37:30] Pique: I will separate dropbox sync from management of RO metadata, or integrate all in the same webapp for management of RO (dLibra?)

[13:38:40 | Edited 13:39:44] Jiten Bhagat: @Pique - we're not sure about this yet. There seems to be a definite need to have some kind of UI for working with ROs

[13:39:07] Jiten Bhagat: but we're not yet sure where this will live, or indeed if it's just one place or multiple places (depending on what interfaces users work with)

[13:39:18] Jiten Bhagat: i suppose that's the point of ROs - the portability aspect

[13:40:32] Jiten Bhagat: (current status: registering of these RO locations is erroring because the backend dlibra is not currently available)

[13:40:43] Pique: ok, what I mean is that if we have a single entry point for RO management is better than having to deal with different webapps

[13:40:46] Jiten Bhagat: (once that is back up we can continue, sorry about this)

[13:40:57] Jiten Bhagat: @Pique - that's a fair point!

[13:41:36] Stian Soiland-Reyes: so I guess the dropbox-sync would just become a particular section within the final interface

[13:41:48] Stian Soiland-Reyes: and in theory once you've set it up, you don't need to go to those admin-pages again

[13:41:57] Stian Soiland-Reyes: unless you reorganize your dropbox folders

[13:42:25] Marco Roos: We are looking at the Dashboard now.

[13:43:03] Marco Roos: First impressions:

Hard to know what to do. Terminology is a bit cryptic.

Can we create containers here?

[13:43:15] Jiten Bhagat: agreed it's confusing

[13:43:46] Jiten Bhagat: clarification is required

"RO container" is the same as "ROs location", i forgot to change all occurences

[13:44:11] Jiten Bhagat: you can create/register containers by clicking on the "Add ROs location" tab on the left hand side

[13:44:16] Jiten Bhagat: (this is what is currently breaking)

[13:44:23] Pique: I guess what is inside containers are different versions of the same RO ?

[13:44:28] Marco Roos: We have tried to create a 'RO location'

[13:44:36] Jiten Bhagat: and is where i need to register the AstroROs and BioROs locations (which will be a user task too)

[13:44:39] Marco Roos: 'Something went wrong'

[13:45:04] Jiten Bhagat: @Marco - yeap, due to the inaccessible dlibra backend :(

[13:45:47] Jiten Bhagat: @Pique - not quite (and i can see the confusion there)... if we ignore the term "RO container" and just talk about "ROs location"... any sub folder within an ROs location is treated as an individual research object

[13:46:20] Jiten Bhagat: (btw, please assume that everything i say is a user task/capability, unless i specify otherwise)

[13:46:51] Stian Soiland-Reyes: What if we call it "My research objects" ? :)

[13:47:12] Stian Soiland-Reyes: like "My pictures"

[13:47:20] Stian Soiland-Reyes: which doesn't have just a single picture

[13:47:26] Marco Roos: Kostas assumes you do all this because DropBox does not support plugin functionality?

[13:47:42] Stian Soiland-Reyes: also to get it done offline

[13:48:04] Marco Roos: NB Kostas may come up with more technical questions by e-mail after this.

[13:48:11] Stian Soiland-Reyes: so if you have more than one computer when using Dropbox, you don't care from which you add the RO files

[13:48:22] Stian Soiland-Reyes: let the techie questions flow!

[13:48:35] Jiten Bhagat: @stian - there is a difference between the list of your research objects, vs all the different locations in dropbox folders where ROs are being deposited, but i do agree that we need a "My Research Objects" section

[13:49:06] Jiten Bhagat: @Marco - not sure i understand the Dropbox plugin functionality question; we are just using Dropbox as one means of depositing content that we can fetch

[13:49:24] Stian Soiland-Reyes: "Synchronize my research objects to..."  - perhaps we could show an example folder listing there. It's a bit hard to imagine now while it's not syncing

[13:49:47] Jiten Bhagat: I have fixed terminology on dashboard for ROs location

[13:49:59] Jiten Bhagat: yes

[13:51:10] Jiten Bhagat: in later work on this prototype, we want to make the UI of robox distinguish between Dropbox related "admin" tasks versus browsing and managing your set of research objects

[13:51:16] Stian Soiland-Reyes: so if you register the folder "Research" as your "ROs location" (or whatever we call it) - then in Dropbox you should see a new folder "Research".   If you now inside there make a new folder called "experiment5" - then you are creating the research object "experiment5" - any files inside Research/experiment5 will be added to the equivalent research object.

[13:51:29] Stian Soiland-Reyes: and then the manifest.rdf from dLibra will be pushed back into Research/experiment5/manifest.rdf

[13:51:44] Jiten Bhagat: yeap, we're in dire need of a concrete example here :)

[13:52:13] Pique: a RO links to a only one path in doropbox, isn't it ?

[13:52:14] Stian Soiland-Reyes: so it could be a workflow, some data files, a few PDFs of papers

[13:53:48] Jiten Bhagat: ( brb )

[13:53:54] Stian Soiland-Reyes: Pique - yes, a particular RO (like "experiment5" in this case) will be represented by the particular path "Research/experiment5" in your dropbox   - but the links within the research object manifest refer to the dLibra URIs - not to Dropbox  (which don't have an URI scheme)

[13:56:16] Stian Soiland-Reyes: I tried to cheat, but Jits told me not to, so we're waiting for Raul :)

[13:57:21] Pique: something I dont understand well is the granularity

[13:57:35] Pique: why different ROs Locations

[13:57:45] Stian Soiland-Reyes: it used to be just one, fixed one

[13:57:51] Jiten Bhagat: this is just in case people use different folders in their dropbox for different ROs

[13:57:59] Stian Soiland-Reyes: but since dropbox APIs don't allow us multiple accoutns while were developing then we had to do multiple locations

[13:58:02] Stian Soiland-Reyes: otherwise it's only one at a time

[13:58:20] Stian Soiland-Reyes: it should be enough with just one folder

[13:58:21] Jiten Bhagat: a good example is right now - we have shared two different ROs locations with two different groups of people from one account (the wf4ever test account)

[13:58:39] Jiten Bhagat: but i guess the usual case is one location for all ROs

[13:58:56] Stian Soiland-Reyes: in dLibra this maps to a "Workspace" which is almost like a user home directory

[13:58:57] Jiten Bhagat: (and as Stian points out, we actually had a technical constraint that resulted in this)

[13:59:21] Marco Roos: I have just asked Eleni to create a table of contents of the experiment she is working on now. I guess that should be(come) an example RO, not?

[14:00:25] Marco Roos: What I think I miss on the dashboard etc, is metadata: descriptions etc.

[14:01:14] Marco Roos: dcterms:description  is empty in the manifest.rdf: would that show up somewhere if it contained something?

[14:01:16] Jiten Bhagat: that would be very useful!, thanks marco

[14:01:52] Jiten Bhagat: @Marco - yeap, that is exactly what is missing from the dashboard, and we hope to have that in for the demo, after getting your feedback now on how you would like to see it done

[14:02:09] Pique: dropbox allow to share folder with other users, what if we provide a shared folder as an RO location ?

[14:02:17] Jiten Bhagat: we anticipate some kind of form to edit metadata about an RO, and then you can view this info in the dadhboard

[14:02:35] Pique: will other users modify my RO ?

[14:03:02] Jiten Bhagat: @Pique - that would work well, especially for collaboration. In cases like those, only one person needs to register that as part of their dropbox account on Robox, and all changes to the ROs will be synced

[14:03:32 | Edited 14:03:34] Jiten Bhagat: yes, in those cases other users who have access to that shared folder on dropbox can modify the ROs in that folder

[14:04:00] Marco Roos: When Eleni is making this table of content we will think of one, but probably several, 'description' columns.

[14:04:34] Marco Roos: We imagine a note about what e.g. a datafile represents, what version it is, what type of document, but also what role it plays in the experiment.

[14:04:36] Jiten Bhagat: ok, sounds good

[14:04:41] Jiten Bhagat: yeah

[14:04:53] Jiten Bhagat: one thing you've touched on there is the relationships between items

[14:05:02] Marco Roos: Will it be possible to make this user-extendible?

[14:05:06] Stian Soiland-Reyes: @Marco - would you be expecting to fill in this information in the web app or through some Dropbox files?

[14:05:12] Jiten Bhagat: that is something that manifest is designed to cater for, but we would need to build interfaces over it

[14:05:35] Stian Soiland-Reyes: I mean, you could in theory be editing the RDF, but that's not exactly a UI

[14:05:42] Jiten Bhagat: @Marco - we would definitely make this user extensible! so allow users to define whatever metadata they want

[14:06:02] Jiten Bhagat: @Stian - how dare you suggest such a thing to users!! ;)

[14:06:19] Marco Roos: @Stian: good question. At least in the dashboard, but I like the idea of providing metadata when and where it is most appropriate; e.g. it could be in Taverna, and showing up in the RO dashboard

[14:07:25] Stian Soiland-Reyes: mm.. so say Taverna should be able to do so - perhaps through a 'magical file'

[14:07:42] Pique: what about importing others ROs ?

[14:07:52] Jiten Bhagat: as an aside - there is a technical issue with multiple apps/interfaces all modifying the same manifest... that can be a synchronisation nightmare! :(

[14:08:14] Stian Soiland-Reyes: mm, good question - so how would you do that? Just drag it in, or browse them and link it up, somehow? We've not addressed ROs within ROs yet.

[14:08:40] Stian Soiland-Reyes: perhaps the question is how could you find the other ROs

[14:09:49] Pique: in the final webapp (robox+dlibra) should be a link to "get me this RO" when you found one you're intereseted in

[14:09:58] Stian Soiland-Reyes: if you just copied it in, it should be possible to recognize the existing manifest and do a link - but it raises the question of where that RO live.. what happens if you change your friends RO

[14:10:21] Stian Soiland-Reyes: Pique, sounds good - and then it should just appear in your dropbox folder?

[14:10:46] Pique: yes, it should be easy for the user

[14:10:47] Stian Soiland-Reyes: you would have it right away and could start opening spreadsheets or whatever is inside

[14:11:11] Jiten Bhagat: yeah

[14:11:23] Jiten Bhagat: this feeds in the "cloning vs forking" discussion on ROs

[14:11:37] Stian Soiland-Reyes: do we need to distinguish between 'my ROs' and 'their ROs' then? In software development we have something called 'forking' when you take someone's code and start tweaking it - should that be allowed?

[14:12:09] Pique: that's whta I call reuse of RO ;)

[14:12:38] Stian Soiland-Reyes: good, good, good!

[14:12:54] Stian Soiland-Reyes: and then it will be linked back to the original RO, so you can be notified or something if there's a new version

[14:13:04] Stian Soiland-Reyes: you might or might not want to get the new changes from the original author

[14:13:49] Pique: yes, that's a good point, "updates are available for this RO"

[14:14:59] Marco Roos: shall we, to begin with, live with the DropBox approach to sharing (e.g. once you share, you make other's owner)? to prevent making things very difficult at first?

[14:15:33] Stian Soiland-Reyes: so here's a scenario, you've found two such ROs which you have tweaked a bit ("reuse", "repurpose"), perhaps it was one workflow and one dataset.

Now you are forming a new RO of the combination of the two - how would you expect that to appear in your dropbox folders?  

Just flat list of research/dataset/, research/workflow/ and research/newstuff/  (where newstuff just has links to the other two) - or perhaps hierarchical, research/newstuff/dataset/  etc

[14:16:00] Jiten Bhagat:  @Marco - i agree with that - start off with something simple

[14:16:18] Marco Roos: Re: linking. I think we need to exploit RDF, linking and extending is its strength. Maybe think of an RDF database to do that?

[14:16:42] Marco Roos: Where is manifest.rdf coming from now?

[14:17:11] Stian Soiland-Reyes: Marco - yes, it's a question if you get write-back to the original RO (like just a shared folder), a copy (ignoring future changes upstream) or some kind of one-way tracking copy (following his later changes), but with possibility to *suggest* the owner to pull down *your* changes

[14:17:22] Stian Soiland-Reyes: @Marco - it's generated by dLibra

[14:17:54] Stian Soiland-Reyes: it's got two parts - the metadata like dc:description which the user can extend arbitrarily (as far as I understand) and the generated listing of all the resources within the RO  (basically a file listing using ore:aggregates)

[14:18:17] Stian Soiland-Reyes: to modify the last list you simply add or remove the files from the folders

[14:18:40] Jiten Bhagat: (in the meantime, feel free to start populating your ROs folders with some example/test ROs, then when the system is back up it will automatically synchronise these and create RO manifests)

[14:19:10] Stian Soiland-Reyes: it's up now! Perhaps we can try.

[14:19:15] Marco Roos: Is that with the wf4ever account? Or our own?

[14:19:49] Stian Soiland-Reyes: either - Jits shared the folders AstroROs and BioROs with  you

[14:20:10] Jiten Bhagat: yeap

[14:20:17] Jiten Bhagat: shared, and added to Robox :)

[14:20:23] Jiten Bhagat: lets see if the sync stuff works now

[14:20:38] Stian Soiland-Reyes: so http://robox.wf4ever-project.org/dashboard/sync_status shoudl say that it's "Not synced yet"

[14:20:41] Jiten Bhagat: @Marco - now you just need to add content on your desktop to teh folder i have shared with you

[14:21:03] Stian Soiland-Reyes: oh, it's PENDING now..  guess the machine is working then

[14:22:17] Jiten Bhagat: yeap, its set to run every 2 minutes

[14:22:55] Pique: I add an RO Location KPMap, should I have said /AstroROs/KPMap ?

[14:23:06] Stian Soiland-Reyes: no, you don't need to register each RO

[14:23:18] Stian Soiland-Reyes: you just go to the shared AstroROs folder on your machine, and do New folder -> KPMap

[14:23:34] Marco Roos: We also added one: epiGeniusExperiment1

[14:23:57] Stian Soiland-Reyes: looks like we confused you both!

[14:24:22] Stian Soiland-Reyes: I guess you both meant "epiGeniusExperiment1" and "KPMap" as two research objects, right?

[14:24:29] Pique: yes

[14:24:43] Marco Roos: I now added that folder within DropBox folder 'BioROs'

[14:24:55] Stian Soiland-Reyes: good

[14:25:00] Stian Soiland-Reyes: and have a look, did you get a manifest already?

[14:25:10] Marco Roos: I'll remove the other later; first see what happens.

[14:25:42] Pique: no manifest so far

[14:25:50] Jiten Bhagat: (i'll remove those locations on robox)

[14:26:22] Stian Soiland-Reyes: seems AstroRS crashed something on our side ;)

[14:26:29] Stian Soiland-Reyes: good test

[14:27:10] Marco Roos: We have some terminology confusion here:

Is a Research Object a container, or is it objects within a container?

[14:27:36] Marco Roos: Kostas and I assume the first, Eleni thought the 2nd.

[14:27:37] Pique: so, if I understand well this is the procedure:

[14:27:41] Stian Soiland-Reyes: you have a container of Research Objects  (like "AstroROs") - each of which is itself a container

[14:28:04] Pique: 1. put the RO on local Dropbox folder

[14:28:10] Marco Roos: Re: confusion. We should make sure there is a help that explains what we mean, preferably with picture and example.

[14:28:16] Pique: 2. declare the RO in the RObox and sync

[14:28:27] Stian Soiland-Reyes: no, you don't need step #2

[14:28:28] Stian Soiland-Reyes: there's just step 1

[14:28:55] Stian Soiland-Reyes: but only if you put it in the *specific* Dropbox folder which you have registered with robox

[14:28:59] Stian Soiland-Reyes: (we won't go peek at your other files)

[14:29:21] Stian Soiland-Reyes: so once - you go and register that particular folder which you want us to watch

[14:29:25] Stian Soiland-Reyes: then you put stuff in there, and that's it

[14:29:32] Stian Soiland-Reyes: you are now creating research objets

[14:29:35] Jiten Bhagat: i think we need to differentiate between "locations" of ROs and ROs - locations are places where multiple ROs may exist within it

[14:29:45] Pique: ah, ok, so AstroROs was already declared, but not by me  ;)

[14:29:46] Jiten Bhagat: you only need to register a location in Robox

[14:29:55] Stian Soiland-Reyes: Pique - yes, unfortunately Jits had to cheat a bit

[14:30:10] Jiten Bhagat: @Pique - yeap, that is the folder i shared with you, via the test wf4ever dropbox account (just for the purposes of this demo)

[14:30:38] Stian Soiland-Reyes: to avoid you having to change your local dropbox installation to connect as wf4ever@jits.co.uk

[14:30:47] Jiten Bhagat: (because we're in "sandbox" mode and only able to use one dropbxo account with the robox app [restrictions by dropbox, temporarily])

[14:31:01] Stian Soiland-Reyes: but once we get our app 'approved' then any Dropbox account can go and register their folder

[14:31:31] Jiten Bhagat: in a way, this does help us test out dropbox's shared folders capabilities with ROs, and brings about questions about collaborating with ROs

[14:32:09] Pique: ok, so the procedure is just well written in the getting started info page of the robox

[14:33:00] Stian Soiland-Reyes: yeah, those would all be empty the very first time. Jits - perhaps we can flush the DB so Pique can pretend he's a fresh user on robox?

[14:33:42] Jiten Bhagat: do you mean delete everything and start again from scratch?

[14:33:52] Stian Soiland-Reyes: or let Marco play a bit first

[14:34:03] Jiten Bhagat: ok

[14:34:21] Pique: can I remove RO from the RObox ?

[14:34:31] Stian Soiland-Reyes: we'll need to investigate which of Pique's files got what to fall over

[14:34:44] Stian Soiland-Reyes: yes, but we don't properly propagate it yet - so it would actually still exist in dlibra

[14:35:01] Stian Soiland-Reyes: (although there's no other way to tell, so for practical purposes it's gone)

[14:35:30] Stian Soiland-Reyes: I guess the researchers would scratch their head as to what 'deleting an RO' would actually mean..

[14:36:00] Jiten Bhagat: yeah, it may be that a repository like dlibra doesn't actually delete stuff, only marks it as "deleted"

[14:36:44] Pique: it's also a matter of organizig your dashboard, that maybe many failed tests of ROs that you would like to remove

[14:37:13] Stian Soiland-Reyes: but what if you put something in there by accident that you really don't want to share or dlibra to preserve in any way - like a.. personal picture

[14:38:18] Pique: yes, or data you do not want to share due to policies

[14:39:08] Marco Roos: I would not worry to much about this right now, and go with DropBox. People have mistakenly deleted whole folders, and then it is very nice that it keeps versions. That would be all I need to start with.

[14:40:23] Pique: ok, I guess that while you're editing/adding meta-data to your ROs you would also like to have the option to remove ROs (and also to remove meta-data if they are extensible)

[14:41:46] Jiten Bhagat: makes sense

[14:42:02] Jiten Bhagat: so a "remove permenantly" option

[14:42:37] Jiten Bhagat: so, looks like some of the files in Pique's RO is causing timeouts during sync, but it has succeeded in syncing a couple of times

[14:42:46] Jiten Bhagat: Marco, can you see your Manifest?

[14:43:00] Marco Roos: yes :)

[14:43:06] Pique: btw, do you want me to remove my KPMap ? :)

[14:43:49] Jiten Bhagat: @Pique, thats fine, it will help us debug the timeout issues

[14:44:09] Jiten Bhagat: as long as you're okay with keeping that data in our test account for a little while :)

[14:44:20] Jiten Bhagat: @Marco - woohoo! something works ;)

[14:45:05] Marco Roos: and I've changed the manifest.rdf, wonder what happens.

[14:45:54] Jiten Bhagat: at the moment we don't sync this back with dlibra

[14:45:56] Jiten Bhagat: it gets overwritten

[14:46:17] Jiten Bhagat: (to avoid the major sync/merge issues that could crop up with many things updating the same manifest)

[14:46:58] Marco Roos: Ok, user request: I think it should contain the creator, and I added Eleni and Kostas as contributors.

[14:47:06] Jiten Bhagat: ok

[14:47:30] Jiten Bhagat: (the RO manifest model/design we're using at the moment is just something pulled from ADMIRAL... once the RO task force has a working model we'll use that)

[14:47:42] Marco Roos: Basically, you would expect that you can provide the information for the manifest somewhere.

[14:47:47] Jiten Bhagat: yeah

[14:47:50] Jiten Bhagat: makes sense :)

[14:48:09] Marco Roos: The dcterms would not change I think; these have become very standard.

[14:48:11] Jiten Bhagat: would you expect this as part of the Robox dashboard?

[14:49:13] Jiten Bhagat: btw, if you have a look at the manifests, you can see that it contains links to the files deposited into dlibra, eg see: AstroROs/KPMap/manifest.rdf

[14:49:23] Marco Roos: See them at least, and then edit them probably. In this case, for these kind of metadata, I think yes: it is the type of information you probably add upon submission

[14:50:00] Jiten Bhagat: ok

[14:50:32] Stian Soiland-Reyes: Marco - you've not yet added anything to epiGeniusExperiment1? :)

[14:52:24] Pique: have added a location Kp in the RObox, I think you should have it in your dropbox, because it has not been declared as AstroROs/kp

[14:52:25] Marco Roos: No, its Eleni's experiment ;)

[14:53:29] Stian Soiland-Reyes: this is why it failed

[14:53:32] Stian Soiland-Reyes: <dcterms:creator rdf:datatype="http://www.w3.org/2001/XMLSchema#string">Marco Roos</>

<dcterms:contributor rdf:datatype="http://www.w3.org/2001/XMLSchema#string">Eleni Mina</>

<dcterms:contributor rdf:datatype="http://www.w3.org/2001/XMLSchema#string">Kostas Karasavvas</>

[14:53:46] Stian Soiland-Reyes: invalid XML.. and we don't push back changes to manifest.rdf yet, as Jits said

[14:53:56 | Edited 14:54:06] Marco Roos: now we did.

[14:54:14] Jiten Bhagat: Pique - you don't need to add any other locations to robox!

[14:54:15] Marco Roos: (add stuff to the folder)

[14:54:16] Jiten Bhagat: :)

[14:54:31] Marco Roos: I have another meeting in 5 minutes.

[14:54:32] Jiten Bhagat: just add new folders within AstroROs

[14:55:01] Marco Roos: Eleni will try to break the system in the coming days ;)

[14:55:01] Stian Soiland-Reyes: Jiten - I belive when we get this live we should remove the "Add ROs location" and only allow one folder

[14:55:10] Pique: @jits yes, I've seen that, have more folders in AstroROs

[14:55:17] Stian Soiland-Reyes: and perhaps we can hard-code it to be "Research objects"  ?

[14:55:31] Pique: they syncronize well

[14:55:32] Jiten Bhagat: @Stian - yeap, agree, it should be an "advanced option"

[14:55:44] Marco Roos: Re

[14:55:45] Stian Soiland-Reyes: Pique, great!

[14:55:45 | Edited 14:55:50] Jiten Bhagat: great, thanks Pique!

[14:55:55] Stian Soiland-Reyes: seems it managed to crunch through KPMap as well

[14:55:55] Jiten Bhagat: Marco - thanks for your time

[14:56:06] Stian Soiland-Reyes: I clicked on a PNG - it has real science!

[14:56:09] Marco Roos: Re: file links in manifest. I tried one and I need a password. I am curious what I would see.

[14:56:11] Jiten Bhagat: this is an initial glimpse into what we can do... we're hoping to come up with a nice demo for the plenary

[14:56:16] Marco Roos: Is dLibra a semantic web thing?

[14:56:36] Stian Soiland-Reyes: should be

[14:56:53] Stian Soiland-Reyes: there would be boring metadata about each file on that link (file size, hash, content type)

[14:57:03] Stian Soiland-Reyes: but the password is autogenerated by the robox now..

[14:57:18] Stian Soiland-Reyes: perhaps we can sneak our way in

[14:57:19] Marco Roos: what does that mean?

[14:57:37] Stian Soiland-Reyes: because it's a dLibra REST interface - but it's username/password protected

[14:57:42] Stian Soiland-Reyes: there's one dLibra user per dropbox user

[14:58:05] Stian Soiland-Reyes: no, that's wrong - one dLibra user per "RO location" folder

[14:58:06] Jiten Bhagat: ooooo, REAL science! :)

[14:58:53] Marco Roos: I really have to go. I hope I can make it to the telcon tomorrow (meeting at hospital with SImone in the morning).

[14:59:07] Marco Roos: bye bye!

[14:59:12] Jiten Bhagat: ok, thanks again Marco and Pique for your time, this has been very useful!

[14:59:18] Jiten Bhagat: take care

[14:59:23] Pique: thank you !

[14:59:28 | Edited 14:59:37] Jiten Bhagat: we can discuss some more ideas in tomorrow's User telcon

[14:59:44] Pique: it loos a very promising and useful tool :)

[14:59:50] Stian Soiland-Reyes: Marco - here's username/password for your evil hacking:

[14:59:52] Stian Soiland-Reyes:               :workspace_id => "xxx",

        :workspace_password => "xxxxx",

[15:00:24] Jiten Bhagat: thanks Pique! still very early days... cheers for putting up with our slow and confusing start ;)

[15:00:26] Stian Soiland-Reyes: and we need to get up some UI to edit metadata

[15:00:31] Jiten Bhagat: yeap

[15:00:34] Stian Soiland-Reyes: so that you see something more than just the manifest file appearing!

[15:00:36] Jiten Bhagat: hopefully for next week's demo

[15:00:44] Jiten Bhagat: and list of the RO contents

[15:00:51] Jiten Bhagat: ie: "what's inside this RO"

[15:01:07] Stian Soiland-Reyes: from where you could start with some primitive annotations of the purpose of each file

[15:01:50] Jiten Bhagat: (please note, i am disabling the background sync for now, will have it back up tomorrow for people to play with)

[15:01:51] Pique: does the manifest change everytime there is a new version of the files ?

[15:02:07] Jiten Bhagat: @Pique - not sure actually

[15:02:10] Jiten Bhagat: it shouldnt!

[15:02:32] Jiten Bhagat: as we want versioning to be a user specific thing (ie: user decides when new versions are created)

[15:02:47] Pique: ok, so it works ok :)

[15:02:53] Jiten Bhagat: (y)

[15:03:46] Pique: do we plan to have more RObox telecons as this one ?

[15:04:22] Jiten Bhagat: it certainly would be very useful to!! i'm unavailable from Wed to Fri unfortunately

[15:04:37] Jiten Bhagat: and on monday/tuesday is the plenary

[15:04:54] Pique: ah yes, the plenary !

[15:05:24] Pique: we will have lots of fun and time to play with it !

[15:05:35] Pique: ok

[15:06:10] Jiten Bhagat: hehe

[15:06:16] Pique: FYI: I think manifest.rdf is updated but not it contents

[15:06:24] Jiten Bhagat: hopefully we'll have the RO metadata stuff ready by then

[15:06:32] Jiten Bhagat: and a better and clearer user interface!

[15:06:39] Jiten Bhagat: ah

[15:06:41] Jiten Bhagat: which RO is this for?

[15:07:02] Pique: KpDrop

[15:07:32] Pique: will check more deeply later

[15:07:33] Stian Soiland-Reyes:

<ore:aggregates rdf:resource="http://ivy.man.poznan.pl/rosrs/workspaces/dbox-VfypgM6ASSKBwwUYOrrESA/ROs/KpDrop/v1/GIPSY.LOG"/>

<ore:aggregates rdf:resource="http://ivy.man.poznan.pl/rosrs/workspaces/dbox-VfypgM6ASSKBwwUYOrrESA/ROs/KpDrop/v1/map.py"/>

<ore:aggregates rdf:resource="http://ivy.man.poznan.pl/rosrs/workspaces/dbox-VfypgM6ASSKBwwUYOrrESA/ROs/KpDrop/v1/SCREEN.LOG"/>

[15:07:38] Stian Soiland-Reyes: seems to have detected all 3 files

[15:08:06] Jiten Bhagat: thanks

[15:08:23] Pique: ok, see you tomorrow for the user telecon

[15:08:31] Pique: i have to go now

[15:08:41] Pique: thanks for the demo ;)

[15:09:38] Jiten Bhagat: cheers Pique, catch you tomorrow :)

  • No labels