WEBVTT 00:00.000 --> 00:11.600 I've ever go ahead. This one then is clever online specific. Optimizing architects 00:11.600 --> 00:18.000 and sitting for multi-tenant collaboration. Okay, so I just try and set the groundwork a little 00:18.000 --> 00:25.680 bit first as to how things were before I went near it. So starting on the left, you've got 00:25.680 --> 00:31.280 like a clever online server called WSD and on the right hand side we have two documents that 00:31.280 --> 00:35.440 are open. So you can see the one kit broker and then there's another one kit broker. Each one of 00:35.440 --> 00:42.160 them is a documents. We each document is opened up in a different process. So you have one 00:42.160 --> 00:46.400 or a clever online server. We have two documents open, each document is open in the process. 00:46.960 --> 00:53.520 And then in between we have this far case which purpose is just basically to generate the 00:53.520 --> 00:59.040 kits that documents eventually open up in. So in between this far case, increase the things 00:59.040 --> 01:03.040 that turn into documents. And you can see that as I've just tried documents in use and there's 01:03.040 --> 01:09.280 one kit spare that's waiting for the next document to come along and each one of those is basically 01:09.280 --> 01:14.080 a bar office. And that's one waiting two open open documents. One of the opens up a document that 01:14.080 --> 01:19.920 would be a kit broker for the moment it's waiting for that to happen. So the far kits job is to make 01:19.920 --> 01:25.520 kits that then are basically free office and the open up documents. And there are spares ready to 01:25.520 --> 01:31.360 use. And this is all done for our performance reasons. The bucket does the initial initialization 01:31.360 --> 01:35.840 opens. It's the start-up of free office. And once it's done, the slow start-up, you know 01:35.840 --> 01:40.960 free office as well as the slow start-up. It does that slow start-up. And then because it 01:40.960 --> 01:45.840 is start-up, the other ones are quick to generate. And then because we keep ones ready to use 01:46.000 --> 01:50.880 already, it's even quicker again. So that's where they're starting from. Each one of the 01:50.880 --> 01:56.000 documents that each case is run inside and it's on CH jail which so each one of those documents 01:56.000 --> 02:00.480 is in its own separate little world. And what's notable is that the bucket that starts up, 02:00.480 --> 02:05.760 it starts up in the ordinary environment. And then afterwards when it sparks the individual children 02:05.760 --> 02:11.040 themselves are inside the jail. So the bucket itself is not in the jail because it was in the jail 02:11.040 --> 02:14.800 then the other ones, you know, they're not ready to be in the jail and that's causing a problem. 02:14.800 --> 02:19.920 So you start off not in the jail and then each document is open up in the jail. In the middle 02:19.920 --> 02:24.480 where the document is inside and these things that's modified, we save the document out otherwise 02:24.480 --> 02:28.640 it's been disaster. But everything else inside in the case is thrown away. So if you're inside 02:28.640 --> 02:34.320 in a clever online, you right click in a misspelled world and you say you want to ignore it. 02:34.320 --> 02:38.560 And then when you exit that session and you start a new session, that word will be 02:38.800 --> 02:44.560 misspelled again because there is no storage between a version between 02:45.440 --> 02:49.600 Kate and since it's except for the document itself. So that's your starting place. Everything is 02:49.600 --> 02:56.240 thrown away and what we want to do is to do two things. One of which when we do make changes for 02:56.240 --> 03:00.640 instance to your spending and you save the these words to be remembered for the future, 03:00.640 --> 03:05.840 we want to place the store of us. We also want to bring in some of the things you have in the 03:05.920 --> 03:10.080 yard nudist version. You have these user dictionaries I just mentioned known as the word books 03:10.080 --> 03:15.760 which keep track of your misspellings. You also have auto text where you have these pre-canned 03:15.760 --> 03:21.280 long sequences of text about a picker topic that when you trigger a particular keystroke, 03:21.280 --> 03:25.920 they expand out these words. So classic ones are we are farm letters and you say the same thing 03:25.920 --> 03:32.480 to multiple people or regular occasions. You just type in a little shortcut and you get out this 03:32.560 --> 03:40.640 bioplytic text that matters and then we would also like to have those pro-organization 03:40.640 --> 03:45.520 of pre-users. So if an organization can have like a stock text, stock auto text, 03:45.520 --> 03:49.840 use a parallelization and individual users have their own individual or top of that and that's 03:49.840 --> 03:56.240 obviously useful for spent checking as well. And we want all that to work and we want to be able to 03:56.240 --> 04:02.080 keep this FarCate hierarchy stuff I showed earlier for the optimization is good. So again, 04:02.080 --> 04:07.280 what we're looking at is single clabber online, maybe two different organizations, maybe a doctor's 04:07.280 --> 04:11.840 one or maybe a research one, nothing to do with each other. They're all hosted by the same 04:11.840 --> 04:17.440 you know, host organization that uses one clabber to mix cloud servers and then there's multiple 04:17.440 --> 04:23.920 users in each one. Each organization who want to have different presets, a different set of initial 04:23.920 --> 04:31.280 auto text, things and initial user dictionaries and each user then may want to have additional 04:31.280 --> 04:36.080 ones on top of that. So you've got that hierarchy. So there's a way things change too. So the 04:36.080 --> 04:41.360 original code of the SD, the original bucket is still there and at the very right inside you still 04:41.360 --> 04:47.040 have each document in the case. What in between you have sub-focards and each one of those is 04:48.160 --> 04:53.840 as the same shared configuration for the organization we're talking about. So the organization then 04:53.840 --> 05:00.240 say it has an ex-doubt and hosts the very presets as one, they're as part of the configuration. 05:01.120 --> 05:08.400 They're downloaders automatically for them and then they instantiate inside the sub-focards. So when 05:08.400 --> 05:14.560 you have an individual document that's opened up from the user, the belongs to particular organization, 05:14.560 --> 05:20.400 we have these ready and running already executing pre-populated with the organization's 05:20.400 --> 05:26.480 auto text and the organization's spent check-in dictionaries and then the individual user ones are 05:26.480 --> 05:32.560 installed on the fair right inside in the individual kits. So the user stuff ends up in the kits 05:32.560 --> 05:38.160 in the right inside the organization of stuff ends up here in this new layer of sub-focards, 05:38.160 --> 05:44.080 where everything is pre-populated. Sub-focards unlike a document when it disappears, 05:44.640 --> 05:49.840 when we come to it and we hold onto them for a while in case you need to reopen a new document. 05:50.400 --> 05:55.120 And then this is all configurable and then we cache everything again, configurably afterwards. 05:55.120 --> 06:01.280 So that's, we have the dash at the end. In the most cases, the shared dash is not going to change 06:01.280 --> 06:06.400 for our organization on a regular basis. So if an organization gives us, for our 06:06.400 --> 06:10.080 the whole effect is that we start off with check-in file in for that, has information on 06:10.080 --> 06:14.240 document, we extend it for this additional thing. We start downloading it as soon as 06:14.240 --> 06:20.320 information is coming from. We're hopefully have all that downloaders or pre-cached and advanced, 06:20.320 --> 06:25.840 but if not, when we eventually get to loading up the document, we download the document as always, 06:26.400 --> 06:30.640 and we get to loading the document, we will delay the loading of that document until 06:30.640 --> 06:34.720 and the sub-focards is available. And when the sub-card circuit is available, 06:35.520 --> 06:40.320 it goes ahead and it launches the spare cases per usual, the spare kits available, 06:40.400 --> 06:44.960 document opens and that. The idea of shared users is the same as what you have in 06:44.960 --> 06:49.360 the labor office already, which is already the concept of shared directory, the user directory, 06:49.360 --> 06:53.600 and also the concept of the process of a pre-set directory, which is when you're a first-start 06:53.600 --> 06:58.240 copies, things come. For this all fits in with how labor office works already. 06:59.520 --> 07:08.160 And to, we splice on our new paths for auto-text and user-dictionaries onto the existing 07:08.240 --> 07:15.280 directories we have for that. And also, there's a lot of settings we haven't available. 07:15.280 --> 07:19.680 Towards the browser settings and online, we can also put them in the same storage as we're putting 07:19.680 --> 07:25.600 these auto-text into. And we already have the existing labor office xu file format for 07:25.600 --> 07:31.920 configuration there. So this is also supported for changing settings as well. Just a one example 07:31.920 --> 07:38.960 little kind of practical issue. I mentioned earlier about the forecast and how that starts off 07:38.960 --> 07:44.240 a user environment and the new within the sub-environment is that in that sub-environment the paths 07:44.240 --> 07:51.600 are different than the outside environment. This is not a problem normally until we start using this 07:51.600 --> 08:00.720 and then we're looking for auto-text paths that actually try out to be inside the individual 08:00.800 --> 08:05.120 document. You're still getting the same paths as they were outside in the original market. 08:05.120 --> 08:10.000 And that's a problem because you can't find them. So there's a whole complicated process of where 08:10.000 --> 08:15.040 you have to basically re-initialize things to get the bypass and then there's a kind of a chain of 08:15.040 --> 08:18.880 three or four issues down along. You have to get one thing re-initialize to get another thing 08:18.880 --> 08:23.600 of re-initialize so that when these dictionaries are in the right place, they can be found. 08:23.600 --> 08:31.200 So there's a number of tedious states. But the outcome is that if you're in the version of 08:31.200 --> 08:36.400 Clara Online, a working version of one of these WAPI servers that supports the feature, 08:37.200 --> 08:41.600 then you can add to your individual user dictionary like that. You can exit your document. 08:41.600 --> 08:46.160 The next time you'll provide any document, desettings have been stored and you'll see them again. 08:46.880 --> 08:51.200 And then at the end of all this to be able to administer them inside and leave it off as 08:51.200 --> 08:57.600 a surf you have an options page. The concept of that options page you'll find in the next 08:57.600 --> 09:04.800 hour which document settings pages which has been extended to include settings that are now hosted 09:04.800 --> 09:09.680 from Clara Online itself. So there's flexibility for the future. And I'll just show you what it looks like 09:09.680 --> 09:17.360 at the moment is that you're inside in the settings of the server itself. And the section here 09:17.440 --> 09:22.080 that says author takes to workbook there like that, they have been coming from 09:22.080 --> 09:27.040 Clara Online itself and gives you opportunity to add additional author text. 09:27.040 --> 09:32.080 I delete the existing ones more books and user dictionaries. That's the personal settings for 09:32.080 --> 09:37.200 individual ones and an admin admin user also has very similar settings. It's actually the same 09:37.200 --> 09:43.600 visually that you can use the override the organization ones. Thank you very much. 09:47.360 --> 09:49.360 you