 Let's talk about the question because we developed this SPG for the development of the SPG I.O. It was just three minutes. We understand that it was set by the set. So, one was not so important. And the primary reason is that there are no representations of this part. It describes actually taking a new place existing in the SPG to develop SPG I.O. Mostly because we want to have one. But that's going to lead to this moment. So, first of all, I don't know how fast I will finish. Thanks to T.F. and their donors for sponsoring that stuff. So, I'm just to make sure I don't forget this at the end. There's always the important to have some financing around it. So, this is kind of the SPG I.O. for life changing at all. So, the status of SPG I.O. parts. I've already shown those, yes. It was more than one. Before the change was just two different SPG I.O. parts. So, they were used one way. They were opening an SPG as document. Directly open document. And the other one, when you already had the document open and inserted an SPG that described both triggering SPG I.O.s but different parts. Both were handling different tasks in the sense of user actions and using different methods to report the different results which is of course unexpected for the user. So, problems are important in this sense. The user doesn't understand why the SPG quality is different because the company is not aware that the SPG is opening or inserting. So, it's just one way. I.O. gets the same point as before. And from the other view, of course, the two important to maintain is always a better fit. So, just to show you what I'm talking about, I have prepared a 5.2, where it's not yet, a 6.2, where it's in. And to show you what's happening, I can trust people on this last table putting this document in just a little pocket. And you'll see it's opening really as an SPG document and then it's in 5.2. And the SPG model is similar to the two objects. You can just see that when you start to track the amount. So, if this object is already not working on the SPG, you won't be able to work on the SPG and the original data is gone. And no way to go back if you save it or something. So, the other way around is when you use a job and it will just drop it. So, you'll see it also gets imported, not too bad. And this time, when you use the SPG, it stays a graphic object, it's a copy of the graphic object for which you will play the SPG, but not destroy it. So now, you are still able to use the context menu to save the original SPG. So, that's the main difference. It's over here. So, the two different techniques to be processed are stuff with one, it's open, it's document import, based on the way to work with the F import. So, it's not a bad idea what's involved. Just, we're sitting in the pubs, the regular F import is guaranteed that you can do everything with the document which you do because it's very well-made and doesn't have to go to government. It just creates a laundry on the fly with caski, which is trust me, and puts it in just low and works more or less as I am used to have. So, not too bad, but it was never too complete and it has speed issues. And you can run the products of consistency and you get anything between CODF import and see SPG using it. So, when some changes to F import maybe it hits SPG import without knowing but it's very simple to run the products. And the SPG, I hope, uses a separate SPG dongle tree on demand by the way. It creates a sequence of things. And this is important because for example, SPG sales, we can show some we in Germany don't have the SPG like where it is in our model not yet but we can handle some if we can decompose and draw some and trim some and export to PDF ever since. So, the general flow is more or less deeply there, but we have no eye or no stuff currently in our file format to get some sales of it. So, I'd like to say that since a few years we've been talking into our creations we can already run those. So, another advantage is it keeps the codes for which the data as you just have seen you can save it again and it's only mine without the HP composition more to say a little bit later. So, it really needs to be focused of course not. So, one of the things that we started with was the differences already mentioned how to get the possible SPG in Germany is an interesting question. So, to get that we don't only do product but also an exporter and sell is an SPG exporter in Japan. So, it's very interesting we can select pages or long pages for our interest and it takes the file export to export SPG and assist even uses one page support not only in interest or even in God it was very surprising when I saw that and it's using the somewhat SMG 1.2 version so this is not really officially existing but I don't know who did it at seven times as export but now it's working and it creates an SPG with an opening of the 1000 ID and then it switches with the other keys and music pages so it's somewhat of a I don't know why and it puts a lot of JavaScript into this export which is managing all this stuff. So, how to get rid of one of them to solve the rough plan is of course get rid of the document SPG if not the main basis or the main page set up that do not touch the existing export if anyone wants to do that just talk to me and the capability to create well, interest arguments with support because the old import and opening as document always created just on the outside but of course to turn around with our SPG we should be able to re-open our open related SPGs so when we export the data from the press it should probably be present when we export the product or the open product also so as soon as possible turn around as possible to say as a page size if you have watched before with the 5.2 if you put it in just this way you just get a really good drawing and then the object is created so so page size is just taken from your international settings that must be doing this size of the SPG it keeps the original data mostly everywhere started with offices just to give you the possibility to have so original data not only of vector graphics but also with graphics in these files you can also find something directly in our files in this data folder then of course it can be really good in a country that is always necessary then so next step, get rid of the document SPG it's not really complicated but you can do it systematically so soon as possible it's more more more complicated I think when we start to read great stuff obviously in office it's interesting so I know a lot of settings maybe so I know some tricks to do and let's compile it there is always going to be different things to study about most of the places but that's not enough you also have to go down and use the symbols of the steps you are leading in between the symbols so you don't forget to be an user or just not to straight forward work but you have to be very careful not to break anything so keep the input filter, that's not good then you don't have a basic starting point because the segment we use to test the input filter keeps it filtered so so so also pretty straight forward that's the first few so we use input filter for different things so this is the one more abstraction that is S&G I talked about before created a sequence of derivatives and the new which is good because it's not important but now the encapsulated boxes is a special S&G filter bless you with OST composition there is a sequence of derivatives you got before the reason behind this is this is the point in memory which is in the data so you don't have to be composing many important parts S&G you just do it as a bunch of memory keep it as long as you want for the end session you can just say it so loading is very fast for 7 and 19 minutes really as we are seeing S&G primitive gets a size so that's also interesting because with S&G primitive it can get the range of the primitive but sometimes you can so it needs to be a size and it should ask the specialized S&G primitive that can trust us so all the S&G is still needed so that's nice and based on some the next step is to create a document and maybe for a person so next there is a graphic object which is the first page which is already there we created the document we created a graphic object to size and position it then there is a page size then it takes a page processor to come to not provide a cell in my first price there is not a page processor to follow besides you have to do something to make it so reasonable for the page size then that's important because you have to do the basic stuff you have to do and at this point the S&G is still not integrated that's nice because it speeds up so so beside let's do the same I did with the 5.2 with the 6.2 so when you open this document you see you get an adapted page and the it's just more about the original S&G file the original S&G is still in the way it's a bad object so nothing must get but you have good quality S&G involved and quality is also better but this is a very simple object and I was just using it just for demonstration so then you can insert it in an existing file same as before with this with this all of such a rule you have to define the page with the page size before and the original size of the S&G is as it is shown both ways both ways with the original S&G so then you get another reason for taking S&G alone compared to the other important you can call it the same as for pictures it's just the quality of S&G so it's a good quality you see a lot of bad stuff especially in the S&G so the S&G was just a much, much more improved and it was a little bit faster for me one of the main advantages is that you don't get draw objects but you get terrible things with difficulty so you have much higher import and you can do much more with S&G if you want to continue to use your S&G so the question was at the same time how do we take from the current S&G if it's so impressive but if it was a S&G it won't be exported so there's a doubt so I took a technology of creating S&G and of course there are typical differences which I'm just great for using these nice stuff so this is a S&G to abstraction of the S&G filter detector which does not only detect files S&G files but also is able to check if it was created of course with two filters before we also had two different sets of filters just the one which does a good job so for example it's probably document type this was a bit tricky because if you had to have a document type it was very easy in the S&G tool layer so due to the filter detector it's possible now to enter S&G types S&G type created by the wrong S&G type created by the wrong and if it's not created by us we forget so so using the same type detector for both which is the type so hopefully in office where you set type and stuff like that but you're given which again is an intimidation it's like one of those which comes from which makes one person mess again so it's just necessary taking the stuff to get the document created so how to detect in my case it's a little bit more complicated because of of course it's very easy to manage stuff it's more or less hard-boiled and don't need a little bit S&G type but not really so this was not as simple to detect in most cases it's okay to take the first 500 bytes so to detect the effect of that also besides so you can already check normally if it cannot yet check if it's just being awesome for the first 500 bytes then there are 6 parameters intensification to the minor page but it's possible that's a really nice thing those primitives I don't know and I talked about primitives in the last few meetings so it's the main one we have this room primitive and the room primitive decomposes to its child content and adjusting the room primitive to type something so there's a room primitive for page so everything which comes into page adjusts to the child context of this room and every renderer or processor of primitives can just not know the primitive that's the main reason for the problem for the decompositions decompositions just by default to the child and that's the way everything any time you have to change no not say the use of primitives with all just find the set and what I've done again is set up a general of our better end vector graphics in some graphics it has this missing search in real freedom so normally we should be able to extract that stuff which should get to many dimensions or not only will it take for graphics to be in a certain dimension to express the primitive we have set a fixed data for any primitive page space which we don't really support normally and it is not really used as a UI for example the primitive page that we have here there's a timeline which I switch to the next page of a primitive something because the only way is in the next slide not sure but that's just a problem I want to mention there's some work that's been done in the future and even for vector formats we have this problem for example with the SMG so that's an interesting question because we can use this space as an hc.3 by which in f1 creates 3 and throw out pages we don't need for each with one page we can create on the hc.1.1 then they create pages and on each page they put the result in one page so that will work it's also possible to use this one primitive which is mentioned in the explanation of single-page format so this is the missing part of this project also a very long time have started to use an explanation in the hc.3 so the goal is the same and we need to continue to burn it but we have no time yet to use it so the current import of one of our 90-page exports is still one page but at least you get the SMG in any case very different but not promising so additional benefits of all the stuff is to be fine vector-based import so for this step there was already an item name to be abstracted a little bit more so why not put the same for the G? I made half a GIL plus also a vector for me we can just the same base so I just did that and moved to the whole GIG that we allow to be besides SMG I think it's called GFIO so it's completely removed from the GIL and it's stuck and isolated behind this barrier for more library why not stay for GTI beta time when we see human DNA because some people complain about human DNA but it's always a question if it's just one how to get the sequence of politics of filters no overhead at all so why not also isolate some GTI beta time stuff behind the human DNA is getting there with some primitives instead of supporting it so now the beta time demo is also completely isolated out of the way of other stuff so the vector-based input performance are more or less all unified at least those important ones I'm sure there are different ones which we could handle the same way but of course I'm missing the look at at that time and as a result there's always a sequence of primitives now so to get that old hope which means internal metafile so if you adjust the time to sequence of primitives and if it's a metafile input if it has any metafile input it's a good fit you can adjust between the first primitives if it's a metafile input or it's also an original metafile so you can adjust it's vector from there and go on and so on so if needed for pain, for example, it's not needed but anyway since years now the input of primitives has been no more than that so I'm using a metafile input for the same parts but I see it as a machine primitive so it's losing all the same effects so this leads to states as if we could use this for any change in the future so if people need to have some system like self-pitch which we can adjust the sequence of primitives works so many are just as a filter always start to switch the data in the background and use a metafile or whatever it doesn't have to be maybe we're input goes from the past also of the excessive primitive representation being excessive to reduce the input idea so this is a base in terms of base and to make it for free it's a good idea what you can do so base input with this always is available for a bias experiment you can handle it for a sooner API it's a good experiment for sequence also and we can get primitive already for separatized separatism for any shape for those who are going in shape or those who are going in storage and now additionally we can just use stg parser and cnn parser to also get a sequence of primitives so what you can send to the send is just use its primitive to T-brand or also excessive for price and time to just change into an experiment so all that all this taking together we can write a new stamp as if we acquired a converter which we express in machine so it's all there the only thing that is not there is the details the primitive can be not be possible to be a primitive outside because the base primitive styles can be unmodified so this is just a question of the master if it's ever needed ok so we have something left that's nice so the trick is the right-missing support to get the primitive it's just one hour of time and I hope we can use it in one day maybe I can send a new file input from and then we build it later so it's nice to modify our access to the system just looking at better buffering more general generic buffering and using stg parser to buffer so it would work I said to focus on x-rays it's just easier so those are the last two first and last just to mention them and now let's see for example with cd 6.2 it will not be too fast because I am always using the versions with cd not really sorry for that but for example if you use them you get 3 so just that we know where it is in cdx for example if I have selected a long page now and you can also do it's this page over to it will take your selection and if you make a point to select and you find and scroll so but when you want to do that so say you want to just a long page which I don't know I don't know where I said it I'm sorry I didn't remember the last one but I don't know so to do it's important to get a few long pages I already showed a lot and I'm really not at the end except if you want to I hope you don't for example it's a very quality difference strategy like that's it thank you