 Good morning, ladies and gentlemen. I'm Dr. Giverman from the Griceford University Archives, and I want to give you a short impression of our work with transcribals in the Griceford University Archives and University Library. First of all, I will tell you something about our first ideas and motivation to work with transcribals. Then I give you some impressions of our practical work and experiences. I will show you some results and will say something about our future plans. The initial impulse to work with transcribals was not a scientific challenge or a research project, but was a more practical idea. We came into contact with transcribals in 2015 when we were discussing the direction of our digitization projects. Between 2010 and 2015, we invested a lot of money and resources to make whole series of archival material available online in our digital library. But the ratio of effort and benefits didn't seem to be balanced. At a certain point of the development, we realized that we had a very costly service there which had a practical value for only a small group of users and we provided pictures of manuscripts which could only be deciphered by a few specialists. So archives are specialists in these institutions for information retriever. So what has this to do with the core task of archives, we asked. And departuring from this question prompted us to make fundamental decision of how to proceed with our digitization projects. And this was the moment when we met transcribals. We were initially skeptical, but the vision of an automatic enrichment of digital documents would maybe solve our dilemma and maybe it could give us a part, maybe a part of a new digital strategy. And so we were simply curious and we started our transcribals experiment. What was this experiment? It's basically the conceals protocol collection and transcribals. Our goal was to read a complete series of manuscripts using the HDR and to make these translated manuscripts available as full text, searchable documents and our digital library for everybody. We chose a close series with 70,500 pages and 123 volumes in this time period, basically three writers. And on the screen you can get an impression of the type of our material and the scripture of one of our writers. In 2015, we transcribed the first 110 pages to train our model and we had high experiences. The results were crushing. Our first hidden mark of model provided a character error rate of about 55% and word error rate 55% CER of more than 20%. And it didn't get much better even if we increased the amount of ground truth material. But we didn't give up and after a certain point, I think in the beginning of the end of 2016, we got our first SITLIP technology-based RNN model and after this point, our work with transcribals changed dramatically because the HDR was so good that we could let the HDR work for us. We had only to correct the results and to refeed them into the training of new models and so we came into a cycle that led us to better and better results every month. You can see it. This was a point when the amount of ground truth and the graph of the CER and after this within a few months, we produced an amount of correct full text documents in the same time that we needed for before the whole year. First, we worked exclusively with archivists as transcribers. After this point, we trained students with the first e-learning tools to let them then carry out the corrections and this increased the tempo again today. Our basic ground truth material is about 2,800 or 3,000 pages. With an average CER of about 5% for all writers, but these average results are more important for the technicians than for us. In practical work, we achieve results of 3 to 4% on a page. Our best result last week was 1.2% CER on a page and very seldom we get more than 5% on a page. So, results. Up to now, thanks to Transcribos, we published 20 volumes of the manuscripts online in our digital library and in a way we wanted it at the beginning of the project. My colleague Bruno Blegel will now show you on the screen how these documents are presented in our digital library. The viewer works with a wide range of possibilities. You can do full-text keyword research with highlighting in the document, tagging results, full-text display and you can download the documents as PDF, searchable PDFs and EPUB files if you want to read the documents on your smartphone. So, here you have one of our volumes. And then you can get EPUB or PDF as you want and you can, of course, get the full text online just as you want. And so, every user can read this handwritten text. Yeah. Okay, how did we do this? As a result of many years of commitment with the integration of OCR, you know this, and the viewer, which most digital libraries did, especially within the framework of the UD project, the University Library already had an interface to take over Aalto results in 2012 and 2016, especially for the, for our project, the interface was expanded to take over imports from Transcribos and that's kind of our workflow, preparing the documents with the Gobi workflow software then in Transcribos and then we got the whole material to the digital library. In the way I showed you. In this way, the Aalto results, input images and searchable PDFs from Transcribos are used for the presentation in our digital library. The next steps, first of all, we would like to complete our project, of course. Another idea gets more and more important for us. Now that we are able to convert images into full text and to do this very fast, we imagine what else we could do with the material, of course. So we would like to test the table editor with a series of Android and statistic documents from the 19th century to get then the content exported into a database to work with the material in a more analytic way than it's possible with a full text file. Another idea is that we would like to try other possibilities of text analysis, for example, by distant reading. That's what we do since a couple of weeks. It's a funny game. We try to visualize results from Transcribos. Here I give you an example which compares the use of Latin words in German texts with the material of two of our writers. The tech pie here works basically with the dictionaries of our HDR models and in the center you can see two time periods, two writers. In the center you can basically see the group of words both of the writers are using and the colored areas are the words used only by one of the writers. So in this case we use an app developed at the University of Leipzig by Stefan Jeniger and we will see what comes out with our experiments with this visualization tools. So my time is over. I hope you could get an impression of our work with Transcribos in Greifswald and if you have questions you're welcome. Any questions? So you have three writers, you train one HDR model. For all of them? For all of them? And three separate models for every writer. Does that work better when you train an HDR model? The beginning writer A at the most pages for writer A and writer B or writer C was only a minority and the special model for writer C and B was better than the complete model and we trained last week a complete model for writer B and it seems that this last model we created was brought to a really comparable resource to the special models because now the writers have got more pages for writer B but also more pages and now it seems that the complete model is more, I have to say it's more balanced I would say. Yes. I was just wondering about what point how many words or pages before you started to be able to switch from transcribing each page to merely correcting the errors on the handwritten transcription page. How many, because I've been working on this and I'm not quite there yet and I want to know what point you get to that wonderful moment when you don't have that wonderful moment. Yeah, two points. We had about 800 pages I don't know the number of words now 800 pages transcribed but at the same moment there were two problems the HDR, at the same moment came the HDR model on SITLIP technology it was much more better than the HDR model that we had before so maybe you will reach this point now much more earlier than we did the other point was the segmentation tool the old segmentation tool was not, yeah I was not a fan of it and most of my transcribers preferred to take the baselines handcrafted Here we had the new RNN model and I think here at this point we launched the new segmentation tool and you can see it on this point we had a lot we could very rapidly prepare documents we had a very good model an average CEO of 10% or under 10% and then the graph of the ground tools material or the correct material grows dramatically with the same input of working time Ok many thanks indeed to Dilkun Bruno I'm pleased to introduce Thaig Deelen from the University of Pound Ok thanks so I'm Wout from the University of Antwerp in Belgium and I'm here to talk about how we use or have been collaborating with working with transcribers at the University of Antwerp So who are we? To put a face on one of the user groups that Wout talked about earlier we are a newly found research centre in Antwerp the Antwerp Centre for Digital Humanities and Literary Criticism but it builds on an older tradition of the age research and digital scholarly editing in Antwerp, especially through the Centre for Manuscript Genetics which is also the centre that is using transcribers What do we do? Basically we digitise and transcribe modern manuscripts so draft materials of literary works and use them to reconstruct the writing process, how the author wrote the works that she wrote so I'm mainly interested in the creative process behind these works and as you can see our documents are quite messy and this is an example of Proust Proust's notebook but our big project that we are working on is actually of the manuscripts of Samuel Beckett It's called the Beckett Digital Manuscript Project Basically we are making an addition of all of Samuel Beckett's manuscripts and analysing them to reconstruct their writing processes Beckett wrote a lot of works so we still have a lot of ground to cover but we already have 5 modules up there of manuscripts of novels, short prose poetry and drama and so just to give you a quick view of how we of our addition it's we digitise these manuscripts we try to offer them in high quality images and we offer the transcriptions also of course where you can read these zones in the text or you can read the entire text of this document in one go and so it's a genetic addition so we are interested in the writing process and so there's also a temporal dimension that we want to put into this addition and that's where compare sentences come in we've numbered all of the sentences in our corpus that allows you let's see to click on any sentence and see how it evolves from document to document so that's the addition that we're making of Samuel Beckett's works so you can see for the single sentences how they evolve from the first draft to the translation to transcriptions in typewriters etc and to the published version so how do we use transcribes well we came in contact with transcribes only recently this edition, the first edition was online in 2011 so we already had a lot of transcribed materials and so last year we invited some of the people from transcribes to for a workshop in our summer school and since then we've been working together to build a Beckett model basically for the Beckett HDR model and we did that first by and we have now about 400 pages of ground truth for this for this HDR model and 350 of those are used using the image text tool which was very which we were very impressed by because it allows and will be discussed later today and tomorrow I guess it allows you to map existing transcriptions onto images and use that to train the algorithm and because we already had a lot of transcriptions this was a very efficient way of getting some progress with the HDR so right now with these 400 pages we're at 11% character error rates for our English algorithm and at 18% for the French algorithm we have two because first we tried to work with just a single dictionary for the entire corpus and we got better results when we predetermined what kind of the language that a document is written in so that transcriptions isn't trying to match it with words but in two languages basically how would we like to use transcriptions further well so we are working with modern manuscripts which are constructed for personal use and not for dissemination or publication we're going to have a lot of specific difficulties like very bad handwriting in many cases because it's only the author himself and this case packet who needed to be able to read his handwriting they're also multi-layered because there's a lot of different writing stages in a single document they follow their own internal logic because the author is still reshuffling bits and pieces of text and using metamarks to link between these texts because there are a lot of heavy deletions especially in very early drafts where the author was not yet satisfied with what he was writing so just a quick example this is not Beckett this is a Flemish author but this is a fragment and you can already see some of the writing layers here in different colors of different writing stages and you can see also an example of the metamark the letter A corresponds to a schema and so it's a link between these two documents the plan on the left tells the author where this specific piece of text is supposed to be and we can also see it's a fragment it used to belong to a different document and it's been cut out so this also gives us some idea of how the text was first written down and how it was reshuffled later which is basically what we are trying to do in the first place there are maybe specific problems for modern manuscripts but they are present in a very high frequency in these in these kind of documents and it's aspects that transcribes is not always equipped to deal with yet so we would want to work together on this and so because we have a lot of these materials we wanted to see if we could try and solve this issue so we've worked together on a funding proposal for a project called catch 2020 computer assisted transcription of complex handwriting and for that project we'd like to train and transcribes specifically on these kind of modern manuscripts with their complex and multi layered textual dimensions and high frequency of noise so that it would become better at transcribing and providing TI-XML output and that's important for us that's closer to the way that we would transcribe these documents in the TI-XML at the same time we also want to see if we can help improve the transcription output by making it language aware whether this would help feeding it syntactical and stylistical information for instance by pre-training a language model for the algorithm on a series of published texts by the author and we try to develop these dimensions and time them because we hope that they may reinforce one another that training transcribes how to read the document structurally would also help put its pre-trained linguistic awareness to better use but those are all ideas that we've been talking about and we've we've submitted a funding proposal for but we hope to be able to work on that more in the future so that's it, thanks That's great, thank you Questions or comments? I read that the way you put the text into the internet is that can you elaborate a little bit on the techniques that you use to show it on the website? Yes, so we transcribe the text in TI-XML and we transform it in Nixosalti Cacoon server but we have specifically the schema that we use is you can find here slash should work so you a-host.u-n-p.be slash b-d-m-p for Beckham-Menders-Lomenska project and that's our schema and there you can see how we transcribe all of these things that like the deletions and the additions and the metamarks how we transcribe them and so that people with similar materials can also transcribe these materials in the same way and use similar XSLT transformation sheets to produce similar results basically Two goals in front of you I suppose that the measurements are not yet in public domain how do you deal with the deletions presenting the internet and its origin before? Yes, yes so it's true they are still on the copyright and we have agreements with the Beckett estate but we work with a subscription based model because of that we try to keep the cost comparatively low but we have a lot of we have demos basically available not for free but to see the entire editions there's a fee that pays for the license basically Thank you for the amazing presentation and an incredibly interesting project Fine thing because I used to study with Dr. Martinix at the International Backup Foundation I was just wondering more of a general editing question What do you do with all these doodles? All the paintings on the page Yes Do they come up? Yes, they do come up we basically we also tag them as doodles and then we then we provide some information and then you can also do a search on them I could probably so that's an example so there are really a lot of little doodles in these manuscripts and we try to interpret them and see if they have a link to the text as well and then we provide descriptions of them and you can look them all up in all of the works and look at them Okay, thanks very much We'll close there Thank you for that presentation and for all our speakers we'll be very happy to continue conversations in one level We'll move on now to Matthias Bohny from the Berlin Brandenburg Academy of Sciences and Humanities Thank you My name is Kai-Michel Würzner My colleague Matthias Bohny is over there We are both part of the OCRD Coordination Project The main thing I want to show you today is our approach to compile a large set of ground truth for historical printings As I said we're part of OCRD which is a funding initiative of the German Research Foundation and it deals with improvement of OCR tools for historical printings So you could say it's kind of an abuse of transcripts since we're not dealing with manuscripts or handwritten text The Coordination Project has three stages We identify to-dos and improvement options for OCR tools and from these we developed a call for proposals which has been released in the beginning of this year and right now we're in the final stage of the reviewing and then a number of sub-projects will work on OCR issues and we're going to merge them into a productive workflow in the end For all the sub-projects we're going to have to provide a lot of ground truth materials for training purposes and we have developed some kind of pilot study of doing this within transcribers Okay, so you could call this the big picture so we have various ground truth sources from different text providers and they're containing either text or structural annotations or both and in very different quality and the idea is to put them into transcribers and harmonize them and provide them as ground truth for the project The pilot study now contains 130 documents with about 500 pages and of course if this is going to work productively we have to provide a lot more So this is a quick overview of the two workflows one with existing text and the other with existing structure within transcribers so we're starting with importing the images like this we're using the option of doing this via the private FTP Okay, then you have the image within transcribers We then run fine reader for an initial layout recognition if we have existing text or we import the existing structure, structural information as page XML since this is the import format of transcribers so this is just the screenshot of running the fine reader so we then come up with a initial structural layout here we have to correct this manually to make this ground truth and if we have the structure available we run some external OCR to get an initial text version so this is the corrected version and then critical point we have to copy the text either the existing ground truth or the initial OCR version by copy and paste region by region and then we have to manually correct the text so this is a somewhat naive approach right so this is not very comfortable there would be other options which are also not very comfortable like creating the page XML completely external importing it into transcribers or doing an intermediate export and re-import so everything not very comfortable that's why we have a few desiderata for the specific topic of ground truth creation but first I have to say we want to say that transcribers is a wonderful tool because it has multiple OCR options it has support for polygonal regions which is one of the only few tools available having that specific support it is a collaborative working environment so multiple persons can work on the same text which is also great in our setup with a lot of students assistance and it has TEI export so as I said for the task of ground truth creation we would welcome the OCR application on specific regions of the page so if we have manually created structural or collected structural information and then run the find reader which is still the best option for us in what's available in transcribers for getting text for Gothic script but if you run the find reader OCR all structural information is removed and replaced by the find reader's own structural layout recognition that's a big, big problem what we would also like to have is our dedicated text import functionalities let's say on the paragraph level this would be great and there is the option of importing meds directly into transcribers but the problem is that possibly existing structural annotations from within the libraries which have been created by the librarians they are not imported so you cannot take them into the tool and it would be a low hanging fruit to utilize all the work done at the libraries what would be great of course would be some kind of automatic guiding in the post correction on the text level there are tools available like Pocoto which could be easily included and it has been mentioned in the talk before transcribers has TI import TI export which is great especially in the context of the German text archive which works completely with TI it would also be great to have TI import and this is not yet not yet realized okay some ways of collaboration between the two let's say communities so OCRD world and read world would be to set up a ground truth repository with all the sources which are created by by a lot of scholars in let's say in the world and we have published the instructions for ground truth creation that we are using internally which are available in source at github and in presentation mode at this link you can find in the slides the slides are also on github so you don't have to photograph them or copy them or writing down this link and it would be nice to see your perspective on this ground truth creation task because this is by now driven by the demands of our project but there are of course overlaps between collecting ground truth for handwritten text or for printed text and yet the Transcriber's user documentation it is there but let's say there are options to improve it so maybe one could move from the very large one page wiki this is available right now to a topic oriented documentation which would make the access to specific tasks within Transcriber's a lot easier for the users and for that we propose Dita because we have a very successful use within our own projects and it would of course be nice to organize a documentation source repository so everybody can bring in his ideas or his specific receipts using Transcriber's into the documentation so thank you for your attention thank you and questions and comments please comment it will be a talk from Australia we'll talk about it tomorrow they are following a very similar approach and one of the main ideas is to replace the time with a digital model so that you don't lose the structure but you will talk about it tomorrow and I'm sure it will be interesting for everyone interested in that okay there are no comments or questions thank you very much indeed for your question and now we will sign the triangular letter from University of France University of Oxford and I'm representing a huge group from the interacademic long-term project the Austrian Bible translator the word of God in German I'm going to show you how we implement Transcriber's in our annotated critical hybrid edition responsible for the research task for us at Augsburg University where I'm from lies with the Bavarian Academy of Science and Humanities in Munich and Augsburg office co-operates closely with another research group at the Berlin Brandenburg Academy of Science and Humanities in Berlin the project duration will be 12 years we've now have finished the first one and it's divided into four individual phases of three years and it's based on materials and procedures developed in earlier BATW and BBAW projects as well as their pre-decessors the pre-reformation Bible translations because there are ones of the 14th century the works of the so-called Austrian Bible translator who considers himself a layman occupies a special position with the Old Testament work the Gospel works in the Psalm commentaries the Austrian Bible translator's provides translations and commentaries for a large part of the Bible with a number of additional eschatological and heriological treatises that can be attributed to him his name is derived from the assumed area of his activity the medieval Duchy of Austria where he was situated on the border between the dioceses of Salzburg and Passau where most of the manuscripts were found and where the anonymous did most of his work translating large parts to the German language the conditions of tradition are quite complex as you can see there are numerous manuscripts and fragments of individual parts of the overhanded down some of them are also art historically relevant of the Psalm commentaries alone so far more than 70 textual witnesses are known and we are still finding more and of the Gospel works 27 complete manuscripts and fragments also early parallel versions of some Gospel work writings origin and authorship of which still need to be termined besides the printed edition of the works there are also plans for a digital version and as a first step since it was already spawned by the German research foundation the edition of the Austrian Bible translator's Old Testament works will as it looks like already be published in 2018 and at the same time the teams in Augsburg and Berlin engage in researching, analyzing and interpreting sources in history of the Gospel works and its versions followed by making the different versions of the Psalm commentaries accessible as a next step the hybrid edition this is our concept of the Austrian Bible translator will publish each version of the individual text as a book however the over can only be presented adequately within the framework of a digital edition since it is possible to cover all the material including various texts and the digitized manuscripts with a partially richly illustrations which open a second level next to the text and also furthermore to realize the synopsis of the different versions of the texts the Gospel works which we are currently working exists as it already been said in two versions soon after it was written the text was edited and whether this also can be attributed to the Austrian Bible translator himself or another anonymous cannot be determined yet but the first assumption seemed very probable the interacademic project now searches for other textual witnesses such as manuscripts and fragments publishing and annotating the texts and the 27 textual witnesses known so far are more closely defined and the reference manuscripts of both editorial stages transcribed all in all the text to be critically edited and commented compromises about 1600 manuscript pages and this is what we are doing right now transcribing all those pages of manuscripts in Berlin because of an earlier start and the already long time existing usage of two step this specific software was used to transcribe the reference manuscript of the older version and its other witnesses and this is how two step looks like it gives you the ability of encoding every aspect of the material text and also makes it possible to print the transcript into a pleasing layout in pdf the shortcomings of two step are not discussed here but of course we have decided to use XML according to the TI to realize our hybrid edition so we use the transcripts of the older version and bring them with a style sheet into XML and then working further with them in Augsburg we are working with transcribers and this gives us the ability to work with the huge amount of text this is how it looks like so we first transcribed about 100 pages of our reference manuscript and then we got the HDR which is giving us the ability to work on a high performance level and also add our TI tagging in some way into the already existing markup which is given by Transcriber the encoding we use is formulated in our guidelines in which we have two markup strategies a manuscript base and a work base perspective on the text the work base perspective is structuring the data according to the units of the abstract work that means books, chapters, paragraphs sentences and so on Aspects we can partly capture in the reference manuscript and transcribers and the manuscript in its individual material and visual appearance that is the structure of the single manuscript its layout also the aspects of the hand in which the manuscript is written the here defined markup is nearly completely considered in some ways in our transcribers transcripts and right now we have an error rate of less than 7% in the HDR transcripts and our student research assistants need 30 to 40 minutes that depends on the student to correct and encode a page completely which is quite good I think Our guidelines are according to the TEI and the DTABF which was already mentioned before which is already a selection of the TEI Our aim is not only to create data which is readable by humans and machines likewise but also to produce reproduction orientated research data to make it possible for other projects to adapt our editing concept Finally for the critical editing of both versions we are using the OxygenXML editor the user friendly author mode you may all know we currently try to extend this by developing a framework for it in which additionally a toolbar is provided where the researcher can enter TEI markup with a push of a button and the entire document text can thus be quickly and simply marked up with a TEI conform XML to create the critical text with its four apparatuses The framework is part of a whole package of different combinable software solutions and technologies Together these options create a dictator framework in which our transcriptions and other TEI documents can be easily edited The digital work environment uses XML database XISTDB as the central repository for its documents and the database is installed on a server and it's available online this allows all project members to access one and the same data inventory and thus work collaboratively In the database it is possible to create an update central indexes for the person places and titles we have already in our transcripts documents and while marking up the text one can also index specific words such as a person to a register To this end a function was programmed that generates a list in which the person's name or place can be selected and then to be indexed and also a website can be built for the project based on XIST, XQuery and XSLT and this website can be opened up to the public our research members Finally with the help of Context, a print edition is automatically generated as a PDF from the TEI XML document and each TEI element is given a specific formatting through a configuration file EDIARUM originally was developed by T. Lothar of the BBAW for the critical edition of Friedrich Schleiermacher where the work environment was used and added Schleiermacher's letters lecturers and calendar it is currently used by a couple of different academy projects but must always be customized for the individual project. In the future that is probably six months from now there will be a basic version of EDIARUM EDIARUM basis for everyone who's interested and if you're interested you may subscribe in their mailing list I show the address later our medieval manuscript now represents a challenge which makes even more special adjustments necessary so we are currently working in a cooperation of the Austrian Bible translator and T. Lothar on EDIARUM edition there is a writing mistake on the file which that is one of the aims could be used by similar projects for the critical editing of medieval text and the witnesses in the future and now that was very fascinating thank you for your attention and await your questions thank you very much any questions or comments please I mean I thought a question the relationship between the print edition and the digital edition do you have any sense of the usage in terms of print sales which is the number of people who might have done more documents the print edition is focused solely on the text as it's a work text for the usage in humanities and our digital edition is trying to get out to everyone to the public and if it's possible to download the critical edited text there this is not clear at this stage because of the relation to the publishers that's always a tricky thing we're trying to do that because it would be senseless to show everything up on the web but not the critical text that's a problem very well thank you very much thank you thank you our final presentation is a film as Marcelio Hernandez from Sweden I think he's working with the Polytechnic University of Poland and we have until tomorrow so we can say any comments or questions when he arrives it also means that we'll probably get a coffee on time thank you to all our brand representatives for being so good in sticking to the timeline first of all we would like to thank the organisers of these transcribous user components for allowing us to present in our case experiences of use proposals for development and aspirations for us to become users the pro-Lope research group is partly responsible for the vision of Lope de Vega's theatre Lope de Vega is a Spanish pre-write who lived between 1562 and 1635 he was a very prolific writer in the Europe for the creation and development of dramas in the so-called Spanish Golden Age the most brilliant period in the history of Spanish literature according to his and his contemporaries testimonies Lope may have written up to 1500 or even 2000 plays nowadays of certain authorship have been preserved thanks to they having been transmitted through also graph manuscripts over 40 handwriting copies and most of all printed copies the theatre of the Golden Age and more precisely that of Lope de Vega constitutes a fundamental chapter in the Spanish and European history this entails an enormous wealth of bibliographic heritage that has remained alive for centuries and is still alive in study programmes publishing, enterprise, bookstores and networks as well as in theatre not only in Spain but around the world the theatre of the Golden Age was a cultural mass phenomenon where wealth all over Spain and productions were sold out not only during holidays but also during working days this involved a huge amount of plays and handwriting and printed copies were needed for representation and reading as a result millions of copies were produced between the 16th and the 19th centuries this is why the theatre of the Golden Age and Lope de Vega are so interesting not only because of its intrinsic quality but also because of the sheer massive quantities of documents involved in its transmission the most important collection of this enormous heritage is preserved in the Biblioteca Nacional de España about which we will speak in a minute Rolope has already published on paper more than half of Lope de Vega's preserved work 175 comedies today Rolope is one of the best 100 research groups in Spain in the field of the humanities in 2014 we published an issue of our Rolope de Vega devoted to digital editions but we have been wanting to pursue for some time we have participated in basic XML DEI digital editing projects such as Canon 60 as well as innovative digital critical editions such as La Dama Bova in collaboration with the University of Bologna and the National Library of Spain the Alliance of Rolope with the Biblioteca Nacional de España as well as with other institutions such as the Compañía Nacional de Teatro Clásico Classical the International Company and the Instituto Cervantes an institution similar to the Gulf Institute of Germany is key for the development of our projects all of them are well known and highly respected institutions and our partnership with them will help disseminate their results of our research Rolope entered with as a more partner through its relationship with the Biblioteca Nacional and its relationship in turn with the University of Bologna the Biblioteca Nacional de España preserves more than 30 million documents including the most important classical collection of golden age theater in the world Spanish National Library wants to get involved in tasks and research projects that contribute to the use of its funds available to the public not only at its headquarters in Madrid but also through the Biblioteca Digital Hispanica Hispanic Digital Library more than 3 million users visited the BNE website with almost 6 million visits and visited or consulted almost 41 million pages the Biblioteca Digital Hispanica offers 180,000 digitized documents composed of more than 22 million digitized pages accumulating more than 1 million visits by 587,000 individual users some 500,000 of the documents have been made all this data confront reports of the institution covering the year 2016 but in addition to the data it's important to underline the active policy of involvement of the Biblioteca Nacional de España in research projects in full awareness of its duty as a public institution in the digital age scenario their entire division of digital development beginning with the Biblioteca Digital Hispanica is developing a continuous work of digitization as well as updating the web and the semantic catalogue the NIDATOS the newly created BNE lab is proof of the library involvement and interest in digital development the partnership between Prolofe and Biblioteca Nacional de España is a two-way path at Prolofe we use the library digitized funds for our digital critical edition which will return to the library as pro-resident culturally and scientifically writes material for the Biblioteca Digital Hispanica in addition and here we return to branch privals Prolofe wants in accordance with the Biblioteca Nacional de España to create or develop tools for digital critical editing to do its own editions but also to put those editions in the BNE lab and make them available to all researchers who decide to use them this is the basis of the adaptation and adaptability proposal that Prolofe would like to make regarding the platform and especially the transcribers edit section many of the cure and theater functions and transcribers are of great interest to philological text editor being able to have an image with file manager in combination with XMOTEI editor and all these in a collaborative platform will be ideal for our project and in our workflow and could also be an optimal solution for the realization of digital critical editions based on XMOTEI philological community in general obviously all the utility that transcribers offers for the management of documents of already made transcription with the version control World CDC are of great interest to us and must be said that the Biblioteca Nacional which will be our main image Provega already has experience in the use of the platform the image processing system of the manuscripts layout and presentation for transcription in transcribers are ideal for a good transcription of the text and at the same time allows Provega and any editor to create and make some text file perfectly aligned and connected to each other the manual transcription is safer both because the editor will not escape lines accidentally and because in the output a synoptic visualization of the text and an automatically synchronized image can be presented in this then work with manuscripts or also in our case with printemata implement you on the computer with an image divided into lines underline the baseline some functions image rotation for really some fragments written vertically like the one seen in the sample and a lower section of the screen that is already the XML-TEI editor is the ideal solution let's now focus on the editor one of the difficulties for the first implementation of the edition with XML-TEI are already admitted by the experts it's to have only few editing tools and because of that we have a complex use that is why the customization and the specificity of the editing screen for our critical text editing tasks should work at maximum performance we think it will be important but perhaps not strictly necessary for the editor to be a graphic editor who responds to the witty-wick philosophy what you see is what you get which will undoubtedly be plentiful for our potential collaborators in our case for the theater of the Spanish Valdenades it is theater written in verse that will be with all the sections of text and TEI tasks that this theater includes leads of characters of dramatic personality state indications names of the characters speaking, speech extanzas in conclusion one could transcribes be developed to become a philological editing tool and platform with accommodation into or access from VNLA? two could it be consigned for use by different research groups even if the signature of an remote or some kind of authorization is required? or alternatively three could it be offered in open codes without authorization and develop it and customize it by and for different projects? these were the case it could become a public to my XML TEI editor that would offer an innovative solution in its specificity and modularity for the philological communities that would be approved by the Biblioteca Nacional de España and hosted by VNLA to utilize and reuse the many digitized images of text available through the institution? the first beneficiary would be a very broad philological community but above all the final beneficiary of these scientific and cultural products will be a linguistic and cultural community of more than 500 million native Spanish speakers the largest in the world of their Mandarin Chinese thanks for your attention one short comment from my side because there was this question concerning TEI editor of course TEI plays an important role and that's the reason why we have included the export we also had the request that TEI should maybe be included for import what I can say is concerning this question we said from the very beginning that transcribers will not be a TEI editor so there are editors on the market and I think they do a good job and what we saw from Angela was very much in the direction we are thinking that basic markup is done in transcribers but more specific markup is done somewhere else in the project of course maybe we could find better ways to make the export more flexible, more configurable and all these things are somewhere on the list that's okay but not a real editor I think our main focus is to give you the chance to create text with basic text so that's our idea of course thank you