Sebastian Garde schreef op 6-10-2014 12:08: > > On 05.10.2014 22:58, Bert Verhees wrote: >> >> >> Op zaterdag 4 oktober 2014 heeft Ian McNicoll <ian at mcmi.co.uk >> <mailto:ian at mcmi.co.uk>> het volgende geschreven >> >> >> Sorry for my late reaction, I was very busy with outdoor sports >> events, I hardly did see a computer, which is very special for me. >> >> >> >> Having said that, as a modeller, working with a wide variety of >> clients, the text file paradigm is very helpful. It means I can >> use industry standard approaches such a git/svn etc. It means I >> can pass the files to colleagues easily. I think it is >> interesting that the >> >> >> by the way, I discovered that it is hard to link directly to on >> archetype in CKM, maybe I am doing something wrong. But it would be >> very handy to communicate archetypes, inclusive connection to other >> tooling like mind maps, etc. >> > Bert, you can always go to the Share With Colleague tab for an > archetype (or other resource) and you will see the direct link for it. > You can also select an action (e.g. go to the mindmap directly or the > discussion page). > There is also a webservice to get the archetypes directly. > If you want to go by the archetype id (which might change of course, > pre-publication), you can also use e.g. > http://openehr.org/ckm/#showArchetypeById_openEHR-EHR-OBSERVATION.blood_pressure.v1
very helpful. To bad it does not show up in the URL, but that is indeed a common issue with web-software. >> I don't think git/svn offers some useful functionality in this >> context, having already a good versioning system. But maybe there is >> something I overlook. >> >> There are some disadvantages to the file-paradigm, that is, why I >> guess you keep on needing the "v", to point out where the >> versionnumbering starts, so software can handle it. It is of course >> because Windows as only OS does not allow a colon as legal character >> for a file name, so you need the "v". >> >> And we are used to having the version in the file name. >> >> The namespace is separated from the archetypeId by a colon. I think >> that is a good thing, the namespace should not be part of the >> content-indicating part of the archetypeId. So should the version. >> But I already mentioned that. >> >> The naming of the file, the tooling should be helpful. There are >> several ways to be helpful, this should be decided by the tool builders. >> >> In CKM, it is, because the archetype-download is under a button! it >> is not possible to do "save as", we have the name CKM offers. So >> there is a file name enforced, and because CKM is often the starting >> source from where an archetype enters a company-network, this name >> remains. This encourages the file paradigm, which is in fact >> disadvantageous. >> > For the browsers I know at least, this is a setting of the browser. > You can select if you want the browser to save downloads immediately > to a pre-defined download folder or save it under a folder/name of > your choice. > Depending on the browser, this is somewhere under settings, e.g. > Chrome: Under setting, go to "show advanced settings" and then to the > "Downloads" section. The option is called something like: "Ask where > to save each file before downloading" > I am fairly confident that if we wouldn't suggest anything there as a > filename that people would complain as well. My browser does this when a simple GET always, or it can display it, or when it can't, it offers to save as. And thereby, in a simple GET, I can use my right mouse-click to "save as"-it. Here the right mouse click does not work, because it is not a link but a form-button. In this construction, you can only click it, and before that you must reconfigure the browser, which may be a problem with other web-software. So I would prefer a simple GET-link instead of a form-button which executes an unknown action. Anyway, this should somehow be more clear, because I am a quite experienced computer-user, and I didn't know how to handle this. Anyway, both questions solved. I am also sure too people would complain if you did not offer a filename, but when you offer a filename, it gives responsibility. And in that case, the "v" inside the filename is not a problem, because the version is sticked to the ID, but it is only a filename. In the internal ArchetypeID, which is authoritative, I would prefer not to see the "v", because it is an unnecessary addition, and can be confusing because the "v" is not normal used in semver-version-counting, and it causes extra string parsing. I think, the reason for the "v" in the internal authoritative ID is inherited by the filename restrictions from Windows. Thanks very much anyway, very helpful. Bert -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.openehr.org/pipermail/openehr-clinical_lists.openehr.org/attachments/20141006/df67db3e/attachment.html>
