Editorial note

1The digital edition of the Sauer-Seuffert correspondence reviewed here was migrated into the ÖNB editions infrastructure and published as version 2.0 on 2nd July 2020 – i.e. immediately before the review was published. This review only targets the previous version of the Sauer-Seuffert correspondence edition, which can in part be accessed via the Internet Archive Wayback Machine.¹

2This planned but not specifically dated migration was already mentioned – prospectively and in general – in the review, which had been submitted to the editor in its first version in October 2018. Its publication – after a peer review and revision phase – was designated for the second RIDE issue concerning scholarly editions of correspondence, scheduled for autumn 2019. Unfortunately, the publication process of this RIDE issue was delayed so that this overlap occurred.

3The ÖNB’s goal with the migration² was to preserve the digital edition as good as possible in its original state. Therefore, the transcriptions as well as the encoding were not revised³ and no major new features were added⁴. In fact, two filters in the person index and the search function had to be dropped (the latter only for the time being) due to technical reasons.

4The most important changes result from the use of the new infrastructure: TEI-XML data can now be downloaded correctly per letter and also in bulk via the GitLab instance of the ÖNB⁵, which greatly increases transparency and reusability. Furthermore, the digital facsimiles are now integrated via an IIIF API and all pages have been provided with a citation note. Finally, the design of the Sauer-Seuffert digital edition has been adapted to the generic infrastructure’s look and feel, which also “solved” the issue with an outdated font mentioned in the review.

5Essential parts of the review are still valid and can be verified through the linked archived pages. Therefore, the reviewer and the editor have decided to publish the review nevertheless, without revising it to reflect the new version.

6It is a characteristic of digital editions in general that they can be further developed and improved. In this respect, any review of a digital edition is created with regard to one particular version, which can be altered or replaced in the future. As a result, content, encoding, presentation, and functions – including errors and shortcomings – may or may not change. To the same extent, reviews concerning former versions of digital editions can still partially be applicable. Besides that, reviews of particular digital editions may also involve the discussion of more general methodological problems and solutions, which remain relevant also beyond project cycles.⁶

Introduction

7The names of both August Sauer⁷ and, perhaps to a slightly lesser extent, also of Bernhard Seuffert⁸, are well-known among German studies scholars in general, and in particular among those who study the history of their own discipline. Both are considered Wilhelm Scherer’s⁹ prominent disciples. Sauer (1855–1926), focussing his work mostly on Austrian authors (in the contemporary sense of the Austro-Hungarian Monarchy; most prominently Grillparzer and Stifter), taught at the universities of Lemberg, Graz, and Prague, where he succeeded Jakob Minor¹⁰ Seuffert (1853–1938) first taught at his Würzburg home university, later succeeding Sauer in Graz from 1886. His contribution to the scholarship on both Christoph Martin Wieland and Goethe is widely recognized, as is his special interest in the epistolary novel genre (in conjunction with Scherer’s Berlin successor Erich Schmidt¹¹ Both are part of a group of scholars that essentially ‘invented’ the Neugermanistik/Neuere Deutsche Literatur branch that adopted goals and methods of the medieval literary scholarship that had formed the main base of ‘Deutsche Philologie’.

8The publishing of the correspondence of Sauer and Seuffert (funded by DFG and FWF)¹², which is held in major parts by the Austrian National Library (Österreichisches Literaturarchiv)¹³, the Würzburg State Archive¹⁴, and other archives, among them the Deutsches Literaturarchiv Marbach¹⁵, was the goal of the publication that its editors – Bernhard Fetz, Desiree Hebenstreit, Marcel Illetschko, Hans Harald Müller and Mirko Nottscheid.¹⁶ intended to do in hybrid form (print and web).¹⁷ Furthermore, it is embedded in a larger research undertaking by the Hamburg based editors who published a number of related correspondences in recent years, unveiling a (postal) academic network that revolves around Wilhelm Scherer as the Berlin based central figure of the emerging “Neugermanistik” field.¹⁸ While the Sauer/Seuffert project was, within that larger framework, aimed at providing both a printed edition and a “web platform”, this review only deals with the latter. The extensive publications section lists Müller and Nottscheid (2016) as the most recent publication. At the same time this book is as close as one can get to a printed edition of the Sauer Seuffert correspondence, as it includes select correspondence of Scherer with both Sauer and Seuffert. The printed excerpt edition of the Sauer-Seuffert correspondence is not yet available on the book market.¹⁹ Still, the sheer existence of a publication list related to the project points to a desirable level of reflection and methodological discussion that is not always met when an institution’s ‘digital first’ (or even an institution’s outline of what would become a digital first in hindsight) is published.

9 Sauer-Seuffert deals with a genuine “Gelehrtenbriefwechsel” type of correspondence, depicting an exchange of letters between two important figures involved with the beginnings of the “Neugermanistik” discipline. The web portal was – judging from the project abstract – planned from the beginning to incorporate “electronically accessible transcriptions of the entire correspondence”, although the technical description remains vague in these early traces of the project and the relation between the printed edition and its digital by-product is not explicitly discussed.

10It is important to note that the web platform’s project description is open about its status, even when it comes to the shortcomings. Judging from the reviewer’s correspondence with project team members, there was also little time and funding allocated to the digital part of the project – which is to be taken into account when putting the outcome’s achievements into perspective while still considering as many criteria as applicable out of RIDE’s “Criteria for Reviewing Digital Editions”²⁰ catalogue. Due to the extent of the correspondence, and a self-imposed focus on the technical and infrastructural surroundings in line with said catalogue, this review cannot capture much of the edition’s content, which the distinguished reader is left to uncover on their own…

The implementation

11The project description gives a straightforward introduction into what the project team intended to provide with the “web portal”:

Diese „digitale Edition“ ist weder als umfassende Einzeldarstellung noch als facettenreiche Ergänzung der Druckausgabe zu verstehen. Sie bietet jedoch ansonsten meist nicht zugängliches editorisches Arbeitsmaterial. Sowohl die Reintranskripte der Korrespondenzstücke der Druckausgabe als auch die noch nicht vollständig kollationierten Arbeitstranskripte der übrigen Korrespondenzstücke werden zugänglich gemacht. Außerdem wurden Teile des Textes nach TEI-Standards codiert, um das Editionsprojekt mit den aktuellen Entwicklungen der Digital Humanities zu verbinden.²¹

12The actual ‘correspondence edition’ part of the “web portal” offers exactly what it indicates: Digital imagery (which is not explicitly mentioned in the project abstract, so it seems noteworthy as a really useful addendum that enhances the web platform) and transcriptions of the Sauer-Seuffert correspondence material plus metadata necessary to identify the material’s holding organisation.

correspDesc

13Major effort went into the encoding of the correspondence’s mechanics. A search for a suitable metadata format spawned a collaboration of the Sauer-Seuffert project group with and within the TEI community via the Correspondence Special Interest Group, resulting in the correspDesc effort to create a schema for encoding correspondences (cf. Stadler, Illetschko and Seifert (2016)).²² CorrespDesc has since been added to TEI P5²³ in version 2.8 (see the commit history of the header chapter²⁴), although in a different place within the teiHeader as the Sauer-Seuffert representation of the correspondence suggests. The underlying interchange format developed on top of correspDesc, CMIF – Correspondence Metadata Interchange Format²⁵, is the data source for web services like the correspSearch website²⁶. Of course, Sauer-Seuffert is also indexed and ‘linked’ to the correspSearch data set (click here for an empty example search). During writing this review, hints were placed with the editing team to update the markup to be consistent with the TEI specifications they themselves established. The metadata centred approach of the correspondence greatly facilitates discovering and finding single pieces of correspondence.

Transcription quality

14The transcripts have been made in TEI P5 and later on included into a Drupal-based presentation layer that combines the multiple views: image, transcription (rendered text) and TEI code. All three of them are vertically segmented by the transcribed pages, which gives a good overview of one letter/postcard/telegram. The user interface is clean and useful for the purposes of reading and browsing the letters and their contents (not taking into account headings that contain “ff”, see below).

15The transcripts are not perfect in all cases, which on the one hand is understandable given the demands of a project that ran for ‘only four years’ and had to build its own infrastructure as well on little funding as a side project to a project aimed at a print publication. On the other hand it is an unlucky coincidence to publish uncollated material that in itself deals with the quality of transcriptions: Sauer and Seuffert do intensively discuss their efforts for both their respective reprint series, and also for the Goethe edition they were both contributing to:

Der Werther ist noch nocht [sic!] fertig(Seuffert addressing Sauer, 4. October 1898, https://web.archive.org/web/20191105160442/http:/sauer-seuffert.onb.ac.at/briefwechsel/brief-von-bernhard-seuffert-august-sauer-prag-am-4-oktober-1898-dienstag )

Ich hatte keine Ahnung davon, daß der einzelne Herausgeber sich soweit [v]on den Grundsätzen entfernen dürfe, sonst hätte ich diese Fülle der Apostrophe niemals über den Götztext ausgegossen; Suphans Bedenken sind sind [sic!] sehr gegründet; aber wo ist die Grenze für die Abweichung?(Sauer addressing Seuffert, 29. March 1889, https://web.archive.org/web/20191105160526/http:/sauer-seuffert.onb.ac.at/briefwechsel/brief-von-august-sauer-bernhard-seuffert-graz-am-29-maerz-1889-freitag )

16The project indicates missing collation elegantly by introducing a ‘traffic lights’ system (translations by reviewer):

• red for “raw transcription, text partially tagged”
• yellow for “transcription checked multiply, text partially tagged”
• green for “transcription checked multiply, text completely tagged”²⁷

17Yet, there is currently no public way to see the respective amounts, as the TEI code only refers to one single revisionDesc//@status=”approved” element per item that does not hold this information. It is unclear where the status indicator data is stored (this is information that needs to be available in order to be able to judge the status of the individual transcription, and it is apparently not saved together with the text itself).

18While it may be strategically legitimate to publish preliminary transcriptions and encourage the community to improve their quality over time, Sauer-Seuffert lacks any means to put such a collaborative approach into practice (be it as simple as a public GitHub repository). In addition, the most obvious way to communicate such transcription errors to the project team – via the contact form on the web platform – did not yield any response.

TEI representation and access

19An obstacle to reusing the Sauer-Seuffert documents is the lack of a data dump or an API. One can think of numerous reasons why this could be useful, e.g. to obtain the edition text for further analysis – or even just to transform the correspondence into a handy .epub file to read the sources in the backwoods without an internet connection.²⁸ The sources are not always validating, e.g. some letters’ TEI code is not well-formed (e.g. 1894-01-25 or 1901-11-11, among a few others that took about 2 hours to manually clean up). The TEI data presented is useful yet basic, as it only includes pb, lb, supplied, unclear, hi, damage and two spans within p, div and body elements. Moreover, the TEI transcriptions are hidden in the HTML code served by Drupal, and escaped within a non-displayed div element, which makes it hard to programmatically take a look at the project’s underlying markup. An example for the first line of a ‘TEI file’:

<span style="color: #009900;"><span style="color: #000000; font-weight: bold;">&lt;TEI</span> <span style="color: #000066;">xmlns</span>=<span style="color: #ff0000;">&quot;http://www.tei-c.org/ns/1.0&quot;</span><span style="color: #000000; font-weight: bold;">&gt;</span></span><br />

Example 1

20A couple of useful common TEI elements for manuscript transcription remain unused, among them choice, subst, add, del, note (which is only used as a container for supplied information that there is no facsimile available), and all the important content of msDesc elements: Any textual critique (apart from regular punch holes encoded by supplied/damage) or scientific comment (which is due to be published in the print edition) that might have been part of the edition project is not part of the online presentation layer of Sauer-Seuffert. Both textual critique and comments would be core functions of any scholarly edition (by the standards that editors like August Sauer and Bernhard Seuffert themselves established in a ‘positivist’ phase²⁹). To be fair, the rudimentary markup is clearly mentioned in the project description.

In search for a search function

21 Sauer-Seuffert search functions are not exactly self-explanatory, nor are they documented or offer advanced features. The search results page offers a number of useful facet filters (by type of document, by year, by places of sender/receiver at the time, and by the dates of birth and death of people in the person index), but being undocumented it is hard to grasp what exactly the filters do (try searching for a random stopword!).

22In its current state, the search function is still a functioning alternative to searching through some search engine’s index limiting by site:sauer-seuffert.onb.ac.at for purposes related to the content of the “web platform”, although it is not possible to query it automatically (e.g. via an API). It might suffice for a full text search, but at the same time lacks basic functionality like wildcard search.

Other tools of navigation (person index, subject index, timeline)

23The person index can be browsed by surname (in alphabetical order),³⁰ but also in ascending and descending order of birth and death dates of all indexed people. From there, it seems that authority file records mostly from the GND have been used to construct brief biographies, with added links to ADB/NDB, ÖBL and Wikipedia, using the GND id as a common identifier seemingly in cases where a GND entry existed at the time of creating the person index.³¹

24Currently, another search operation is necessary to find textual mentions of a person mentioned in the index, as there are no backlinks included in situ. This significantly increases the reader’s effort, especially in cases where names are abbreviated or indirectly referenced (“her daughter”) and will not be found (considering that wildcard search is not possible).

25A subject index is definitely not an easy undertaking in general (depending on various variables that are hard to reconcile). Following two bright minds through their correspondence over almost a lifetime hinders building a useful subject index quite a bit. Sauer-Seuffert tackles this challenge by offering what they call “Themenlinien” (thematic lines). That notion for subject categories is likely not by chance very vague – as its implementation is highly subjective and seems to depend on what was interesting to the editors (especially with respect to the Hamburg team’s history of Germanistik approach). The list of subjects – in its entirety – consists of:

• Vierteljahrschrift für Litteraturgeschichte (15)
• Freundschaft Sauer – Seuffert (19)
• Wiener Neudrucke (10)
• Euphorion (27)
• Grillparzerforschung (22)
• Akademischer Werdegang (19)
• Beziehung zu Wilhelm Scherer (9)
• Deutsche Arbeit (4)
• Projekte zur Österreichischen Literatur (15)
• Stifterausgabe (7)
• Deutsche Litteraturdenkmale (12)
• Weimarer Goetheausgabe (8)
• Wielandforschung (19)

The numbers in parentheses refer to a count of correspondence items that apply to the subjects in the whole transcribed corpus – and point to a slight disproportion: only 177 instances of 13 subjects (non-exclusive) would be mentioned in 1248 correspondence items?

26A very brief check revealed that e.g. “Weimarer Goetheausgabe” does not refer to either the Werther ³² volume of the “Weimarer Ausgabe” edited by Seuffert or to the Götz ³³ volume edited by Sauer. In the current state, this subject index does not help much and for sure does not replace a detailed reading of the Sauer-Seuffert correspondence or a full-text search.³⁴

27Following the hint in the project’s final report, one of the project’s major developments is the timeline prominently featured on the start page. This might be one of the major contributions for outreach and popularizing the importance of the whole endeavour, which in turn more than satisfies the notion of a “web platform”. The timeline’s initial view draws from biographical data without direct connection to the correspondence’s texts, but offers access to the topic fragments mentioned above and respective quotations on clicking into the topic bar. This definitely is a nice feature (although user experience-wise, it is easy to miss) as it relates the correspondents’ biographies with some of their letters.

Contact form (and font issues)

28The web platform does feature a contact form³⁵, but upon trying to contact the editors this way there was no reaction (even after talking to two involved people). This might have to do with the web platform being a ‘finished’ project – which in fact is not communicated on the website but might be inferred from the project runtime dates –, and nobody specific is to blame for the fact, but still it leaves a stale impression of an orphaned website. On the other hand the fact may be seen as a reminder for similar projects – and, above all, institutions planning to host DH projects! – to think ahead and plan responsibilities for the time after the funded project phase. There is a definite need not only to keep a service up and running, but also to minimize both technical and editorial errors.

29A design decision lead the Sauer-Seuffert team to use a special font. They selected Born by Carlos de Toro. The reason for the reviewer to contact the type designer was to ask about the strange-looking ff ligature in his font that resembles a double long-s (ſſ) ligature. Of course, this small error gets seen if it is put in the header of a web portal like the Sauer-Seuffert correspondence. A snapshot of the landing page is available here. Carlos de Toro had already spotted the ligature error in Born years ago, but did not get around to fixing it in the public version(s) of the font.³⁶ While the font has been pulled from distribution by its creator since, copies still float around on various (mostly illicit) websites. Some of these violate the font licence that clearly states “You can’t: -Redistribute this typography by other way or different websites (except if you have the author’s permission)”. The project team repeated this mistake by hosting its own copy of Born on their web server. Yet, upon request by the reviewer the font’s creator acknowledged the project’s use of his typeface and granted permission in retrospect.

Is it a Digital Scholarly Edition?

30RIDE’s Criteria for Reviewing Scholarly Digital Editions³⁷ adopt a clear framework for answering that simple question. They list three “necessary conditions”:

• a justification of the editorial method adopted and a clear description of the rules that guided the edition,
• compliance with scholarly requirements towards content and quality, which includes that the self-stated rules are followed,
• and an editorial concept that is not restricted to the technological limitations of print technology but that implements a “digital paradigm”

31Applying this to the web platform yields that:

• Sauer-Seuffert does indeed justify their methods and editorial vision (in the “Zur Edition des Briefwechsels” section on the edition), but they fall behind it in the HTML representation of what might once have been TEI source files.
• The rules from the quoted section are mostly followed, TEI standards only partially. For requirements coming from German studies and/or scholarly editing, the web platform does delegate all questions to the printed edition.
• The third condition is the most questionable for Sauer-Seuffert, as their editorial concept was firstly informed by print technology (the other pillar of the project, which has not yet ripened to fruition), and secondly by technological limitations of a “digital paradigm” not quite in alignment with the digital editions that are sprawling elsewhere. Putting commented-out and escaped XML fragments in HTML served by Drupal is in fact counter-intuitive. Most other DSE use native XML databases (the ‘easy way’). It would of course have been possible to create a TEI-first workflow with Drupal allowing for similar functionalities compared to an XML database, but Sauer-Seuffert went the other way and integrated TEI data as a side effect. Putting out scans and raw transcriptions without either an API for others to build on the material or at least some kind of working feedback mechanism (e.g. crowdsourcing raw transcripts³⁸) would not answer to the specific demands of the particular “digital paradigm” that the Criteria refer to. In contrast, the Sauer-Seuffert timeline does indeed instantiate a “digital paradigm”, even if it only does so for 177 source items.

Good news on ÖNB infrastructure for digital scholarly editions

32The Austrian National Library has already launched their next Digital Edition, the Andreas Okopenko Diary Edition³⁹ (project descriptions: univie⁴⁰ | ONB⁴¹, funded by the FWF⁴²).⁴³ With this edition, the Austrian National Library started an edition platform that should take existing standards and best practices even more seriously. Technically, it is based on the Fedora⁴⁴ Commons implementation by Graz University⁴⁵ (Seuffert’s academic home institution for many years) called GAMS⁴⁶. Efforts to integrate or completely migrate Sauer-Seuffert into this technical infrastructure remain to be seen, with the likely effect of the result being a Digital Scholarly Edition proper. With their GAMS implementation and the Okopenko material, the ONB took a further step in the right direction, as they enable access in a sense even exceeding the understanding of the funding bodies’ (sometimes legally unclear) Open Access policies (FWF recently co-founded cOAlition-s⁴⁷), as they are sharing all TEI XML data in downloadable form under a CC-BY-SA 4.0 licence. This goes hand in hand with the recent convergence of reusable platforms and modules that some Austrian institutions (including the Austrian National Library itself) already do with the Kompetenznetzwerk Digitale Edition (KONDE)⁴⁸ (funded as a “Hochschulraum-Strukturmittel-Projekt”). A big challenge for these undertakings is the integration of e.g. prosopographical data such as places, people and institutions which would facilitate Linked Data integration beyond correspDesc in order to avoid building parallel insular resources. It might still need some time to set up and implement the various components, but the vision is already in place, planted by projects like the Sauer-Seuffert edition.

Conclusion

33All in all, the Sauer-Seuffert correspondence “online platform” (they avoid the term “Digital Edition” for a reason) is one of the project’s most visible deliverables to this day, although it was created only as a spin-off without extra funds. Furthermore, it captures the whole of the available correspondence (approx. 1200 pieces), where the selection to be printed will only cover some 300. As such, the platform is an invaluable addition to the whole field of history of Neugermanistik and saves the correspondence from being buried in the archives again (especially after a select edition became available hardly anybody would have thought about looking at the portion of archive material that its editors left out). On the negative side, the platform suffers from the same problems other “finished” Digital Humanities projects are facing after their funding period: They get abandoned and do not follow up even on full-bodied promises. Also, and beyond this ‘DH problem’ (which, strangely enough, connects to the problem of “connatale Vergreisung” [connatal senescence] mentioned in the title of one of the project’s other publications (Illetschko and Nottscheid 2017) and that originally refers to the aporiae of big scholarly editions as the Weimar Sophienausgabe of Goethe’s works), the printed book featuring a selection of the most interesting parts of the correspondence has been promised to be published “soon” by Hirzel Verlag, but remains to be brought to the light of publication.⁴⁹ On the positive side, the Sauer-Seuffert project had serious impact on the development of the correspDesc metadata set, which is such an important addition to the TEI’s standards that it recently was awarded the Rahtz Prize for TEI Ingenuity 2018 together with CMIF and correspSearch. Judging the ‘web platform’ as an institutional-level Digital Scholarly Edition ‘digital first’ of at least two large institutions – from the distance of six years (the project started in 2012), which is a long period in the quickly changing Digital Humanities field –, it is important to highlight the value and effort that the Sauer-Seuffert project contributed and continues to contribute to the Digital Edition eco-system as a whole.

34The bottom line for research institutions and funding agencies alike would be that proper scholarly editions (this especially applies to digital editions, but holds true for other – even ‘old’ media as well) need proper funding – especially dedicated funding for technical implementations, proper technology, proper oversight, and proper provisions for the time after edition projects are ‘finished’. The separate tasks of curating data and ‘keeping it alive’ both cannot be given over solely to the individual researchers (in times of short term contracts). All of this will not be possible without allocating effort – and funds.