Spielerische Ansätze in sammlungsbasierten Crowdsourcing-Projekten. Impuls beim Gamification-Panel der Österreichischen Citizen Science Konferenz 2018

Während der Österreichischen Citizen Science Konferenz vom 01. bis 03. Februar in Salzburg beleuchtet das Panel „Chocolate Covered Broccoli “ die Anwendbarkeit von Gamification für Citizen Science-Projekte. Das Panel ist als Fishbowl-Diskussion mit drei kurzen Impulsen…

Quelle: https://bkw.hypotheses.org/691

Weiterlesen

FES: Die Bibliothek der Friedrich-Ebert-Stiftung digitalisiert den „Vorwärts“

http://www.vorwaerts.de/artikel/fes-digitalisiert-vorwaerts Der „Vorwärts“ kann auf eine lange Geschichte zurückblicken. Gegründet am 1. Oktober 1876, wurde er 1933 von den Nationalsozialisten verboten. Jetzt sollen alle Ausgaben aus dieser Zeit digitalisiert werden. Dabei wird die Hilfe der Leser benötigt. Via: https://twitter.com/FEShistory/status/682948259762028544

Quelle: http://www.einsichten-online.de/2016/01/6299/

Weiterlesen

Vortrag „Designing Heritage Crowdsourcing Projects“

von Margit Wunsch Gaarmann, FU Berlin

Dr. Mia Ridge (Digital Curator, British Library) wird im Rahmen des Projektes „1914-1918-online. International Encyclopedia of the First World War“ am Montag, 07.12.2015 um 18 Uhr s.t. über das Thema „Designing Heritage Crowdsourcing Projects“ sprechen.

Zum Inhalt: Based on a review of the most successful international crowdsourcing projects, this talk will look at the attributes of successful crowdsourcing projects in cultural heritage, including interface and interaction design, participation in community discussion, and understanding participant motivations.

[...]

Quelle: http://dhd-blog.org/?p=5910

Weiterlesen

Was ist schön? Eine Antwort von den ARTigo-Spielern

schönDie Frage „Was ist schön?“ ist alt und wurde vielfach diskutiert. Nicht nur Kunsthistoriker, auch Psychologen, Evolutionsbiologen und Philosophen trugen und ragen dazu bei, eine Antwort zu finden. Obwohl von vielen Disziplinen aus betrachtet – gibt es keine klare Antwort. Dafür viele Argumente aus zahlreichen Perspektiven.

Eine Perspektive kommt nun hinzu: die Tags der ARTigo-Spieler. Somit handelt es sich hier nicht um eine einzelne oder gar meine persönliche Meinung, sondern um ein crowdgescourctes Geschmacksurteil.

Ich habe spezifische Datensätze selektiert, die die Tags von impressionistischen, expressionistischen und klassizistischen Bildern enthalten. Dabei habe ich festgestellt, dass die Tags „schön“ und „schönheit“ nur bei der Berechnung von Korrelationen auf dem Datensatz, der die Tags der Bilder des Klassizismus enthält, vorkommen. Sicher, es ist nichts Neues, wird doch gerade in dieser Epoche die Schönheit idealisiert. Dennoch: Trotz unserer reichen Bilderwelt und der zeitlichen Distanz zum 18. Jahrhundert empfinden ARTigo-Spieler die Bilder immer noch als „schön“. Und anders als bei Laborstudien, bei welchen Probanden Bilder mit expliziten Fragestellungen vorgelegt werden, wird nach Schönheit beim ARTigospielen gar nicht gefragt. Der Spieler entscheidet selbst, welche Tags er eingibt. „Baum“ und „Himmel“ sind sichtbar vorhandene Objekte. „Schön“ jedoch, spiegelt den Eindruck der gesehenen Objekte im Bild wider, was einer höheren Verarbeitungstiefe entspringt. Und abgesehen davon: Auch Bilder aus anderen Epochen sind schön. Aber schon rein rechnerisch die Bilder des Klassizismus besonders.

Digitale Bildquelle: www.artigo.org
Künstler: Louis Sené, Titel: Dame in Liebesaltar, Ort: Celle, Bomann-Museum, Datierung: 1794-1794
Es war das erste Bild, das mir von ARTigo bei der Suche nach "schön" angezeigt wurde

Quelle: http://games.hypotheses.org/1889

Weiterlesen

Twitter zur Unterstützung von Kooperationen – Kurze Vorstellung auf dem Rheinischen Archivtag 2014

Der Kurznachrichtendienst Twitter bietet trotz der Beschränkung auf nur 140 Zeichen pro Beitrag zahlreiche Möglichkeiten des Informationsaustausches, der Vernetzung, des kollaborativen Arbeitens und der internen und externen Kommunikation. Twitter ist noch dazu im Einstieg äußerst niedrigschwellig, Beiträge lassen sich auch ohne Anmeldung lesen, die Einrichtung eines eigenen Accounts benötigt nur wenige Minuten und es werden auch praktisch keine persönlichen Angaben abgefragt. Die Anwendung ist leicht zu bedienen, und mit nur wenigen Klicks lassen sich sowohl mittels eigener Tweets Informationen verbreiten als auch abrufen.

Tweets enthalten aufgrund ihrer Kürze meist nicht nur Text, sondern häufig Links, Fotos oder Videos und lassen sich mit anderen sozialen Netzwerken wie z.B. Facebook verknüpfen. Besonders schnell und einfach ist auch die Verbreitung von Inhalten über die Funktion des Retweetens. Die Startseite oder Timeline bietet eine Übersicht der aktuellen Beiträge aller Twitter-Accounts, die man abonniert hat, denen man „folgt“, wie es bei Twitter heißt, man kann die Beiträge aber durch das Anlegen von Listen auch filtern. Hier beispielsweise eine Liste mit deutschsprachigen Archiven auf Twitter. Besonders praktisch auch die Funktion des Hashtags, also die Vergabe von Schlagwörtern, über die sich alle Beiträge zu einem speziellen Thema, die mit diesem Schlagwort versehen wurden, aufrufen lassen – hier als Beispiel das Thema Archive und Web 2.0.

Mit den Tweets lässt sich auf weiterführende Inhalte verweisen, über Publikationen wie z.B. Blogeinträge, Vorträge und andere Tätigkeiten oder auch Veranstaltungsprogramme informieren, das eigene Netzwerk ausdehnen und ständiger Kontakt mit den Fachkolleginnen und –kollegen halten. Über Twitter lassen sich außerdem Tagungen und andere Veranstaltungen – häufig nahezu in Echtzeit – verfolgen und kommentieren– vorausgesetzt natürlich, es sitzen twitternde Teilnehmer im Plenum. Hier links zu sehen Beiträge vom diesjährigen Archivtag für Rheinland-Pfalz und das Saarland in Worms. Mittels einer Twitterwall – hier rechts zu sehen bei der diesjährigen Tagung Archive 2.1 in Stuttgart – können die Beiträge allen Teilnehmern vor Ort sichtbar gemacht werden.

Besonders unkompliziert lässt mittels Twitter auch ins Gespräch kommen – der oder die gewünschten Gesprächspartner müssen lediglich mit dem eigenen Account angesprochen bzw. „erwähnt“ werden. Da diese Gespräche öffentlich sind, können sie auch andere an dem Thema Interessierte verfolgen. Der nichtöffentliche Austausch ist aber natürlich auch möglich, dafür gibt es die Funktion der Direktnachricht. Und – wie rechts zu sehen -, werden mittels Twitter sogar Interviews geführt – hier im Vorfeld und als Vorbereitung auf die Tagung Archive 2.1.

Eine besonders schöne Form der Vernetzung sind für mich persönlich auch Twitter-Aktionen wie etwa anlässlich des Internationalen Tags der Archive / AskArchivists Day, bei denen man internationale Archive und deren Bestände kennenlernt – und das kann und soll sogar Spaß machen, wie man beispielsweise an der Aktion #ArchiveShelfie sieht, bei der Archive auf der ganzen Welt die schönsten Aufnahmen aus ihren Magazinen posteten. Die Beiträge von Archiven im Rahmen dieser Aktionen gehen übrigens jeweils mittlerweile in die Hunderte. Entdecken lassen sich andere Accounts auch über den so genannten FollowFriday. Hier werden Accounts über ihre Nutzernamen verbreitet, die man selbst empfehlen kann. Und sogar Crowdsourcing ist möglich, indem man andere Nutzer um Hilfe bittet. Häufig wird hier der Hashtagt #followerpower verwendet.

Quelle: http://archivlinz.hypotheses.org/269

Weiterlesen

Crowdsourcing in Theorie und Praxis

Am Samstag, den 21.6. habe ich an der Evangelischen Akademie Tutzing einen Vortrag  mit dem Titel gehalten: „Wie Crowdsourcing unser Leben und Arbeiten verändert – Chancen und Risiken eines neuen (?) Verfahrens“.

Das Besondere dabei war, dass ich ihn vor den Alumnis des Heidelberger Life-Science Labs gehalten habe. Das Life-Science Lab ist eine Einrichtung des Deutschen Krebsforschungsinstituts. Hier wird der naturwissenschaftlich interessierte und begabte Nachwuchs gefördert. Schüler ab der 8. Klasse bis zum Abi haben die Möglichkeit, sich zu bewerben und dann in Arbeitsgruppen zu forschen. D.h. sie können an Forschung mitarbeiten oder auch ein eigenes Projekt initiieren. Die Alumnis selbst leisten engagierte Netzwerkarbeit um das Life-Science-Lab herum. Und diese jungen Leute habe ich erlebt.

tutzing

Für den Vortrag waren zwei Stunden inklusive Diskussion angesetzt. Erst dachte ich, das ist aber ganz schön reichlich viel Zeit. Begonnen habe ich mit ein paar kleinen Querdenker-Aufgaben, die praktisch in das Thema einführen sollten. Da habe ich gleich gemerkt, dass diese Gruppe sich kennt und gut miteinander arbeitet. Sie legten sofort los und waren im Nu mittendrin:

Nach dem Vortrag hatte ich noch einige Fragen für eine Gruppenarbeit vorbereitet. Das war aber gar nicht mehr nötig, denn die Diskussion kam vortragsbegleitend in Gang. Und das war richtig toll. Die jungen Leute waren interessierte Zuhörer und Diskussionsteilnehmer. Was man manchmal nach Vorträgen erlebt, dass Fragen etwas zäh gestellt werden oder sich Zuhörer durch die Fragen selbst exponieren wollen, war hier definitiv nicht der Fall. Hier waren alle „auf Augenhöhe“. Einfach super!

Ein Teilnehmer bemerkte zum Schluss noch, dass die Gruppe ja selbst Lösungsvorschläge erarbeitet und gesammelt hätte, und dass dies doch auch Crowdsourcing gewesen. „Ja“, konnte ich da nur sagen. Denn wenn man für das Thema sensibilisiert ist, merkt man erstmal, wo Crowdsourcing überall stattfindet und das es letztlich ein Merkmal unserer gesellschaftlichen Entwicklung ist.

Hach!

Quelle: http://games.hypotheses.org/1652

Weiterlesen

Die Crowdsourcing-Plattform Velehanden.nl

Eine der zentralen Fragen bei dem Einsatz von Web 2.0 in und für Archive ist die Frage nach dem Nutzen für die Archive. Die Vorstellung etwas online zu stellen, was dann von anderen erschlossen, bzw. bearebitet wird ist eine der Zukunftsvisionen, die viele Archive verfolgen. Scheinbar bieten sich einige Bestände, wie z.B. Fotobestände für das sogenannte “Crowdsourcing” geradezu an. In den Niederlanden läuft schon seit Jahren erfolgreich die Crowdsourcing-Plattform www.velehande.nl Ellen Fleurbaay und Nelleke van Zeeland vom Stadsarchief Amsterdam berichten hier über das Projekt, die Betreuungsaufwände, den Nutzen und die Ergebnisse bis jetzt. Dieser Beitrag in englischer Sprache bassiert auf einem Vortrag, der beim Deutsch-Niederländischen Archivsymposium im Okober 2013 in Arnheim gehalten wurde.

 

Velehanden.nl: what does it take to make a crowd?

Von Nelleke van Zeeland und Ellen Fleurbaay

10_Zeeland_homepage VHVeleHanden, which can be translated as Many Hands, is the online crowdsourcing platform initiated by the Amsterdam City Archives. It is active for almost two years now, so high time to look back, evaluate and value the results. A short history of the Amsterdam City Archives and our online activities will provide the necessary context and background of this project. Then we will focus on how the VeleHanden project was initiated and organised in a public private partnership. Then we will arrive at the most important question of crowdsourcing: how to create and keep a crowd. In conclusion, we will touch upon the results of VeleHanden. What did we accomplish? How is the quality of the work?

A short history of the Amsterdam City Archives

First of all some basic information on the Amsterdam City Archives. Since 2007 we are located in a beautiful building in the heart of the Amsterdam City Centre. We receive an average number of about 100.000 visitors a year. That is a figure we are very happy with, as before 2007, when we were housed in a picturesque little building on the outskirts of the city, we never attracted much more than a quarter of this number. And even this quarter was declining, in an alarmingly rapid pace. At the same time we were amazed to see how many visitors were attracted to our first website in 1998 and how fast the number of web visitors was growing. Without much effort, in a few years the website attracted ten times more visitors than our reading rooms.

The website started in 1998 and, as most websites in those days, it was a sort of online brochure, offering illustrated stories about highlights in our collections, information on opening hours and other practical things. Then we saw an Image database, probably Getty’s. We fell in love with it and wanted one for ourselves. In 2003 we had our own Beeldbank.

10_Zeeland_projectpage VHAbb1. Homepage + Projectpage

 

 

 

 

 

 

Then some visitors asked: ‘Why an Image database? Your core business is archives, so where are the archives? We had to admit that this was an apt remark and we set ourselves to get an overview and catalogue of our archival collections online. We named it: Archiefbank, or Archives Database. Our colleagues said: well done! But our customers were not satisfied and they asked: ‘Where is the button to see those documents? I cannot find it’.

Again an apt remark we thought. So we decided to start scanning. This time our colleagues laughed at us and asked us if, by the way, we had forgotten to calculate how long it would take to scan all those kilometres of documents. Of course we had not forgotten; it is an easy sum: with an average production of 10.000 scans a week it would take about 420 years to scan our collections. But the question is not about how much documents we have in our collections, the question is how much documents our customers ask from us. And that does not outnumber about 15.000 scans a week. That is why we started a ‘scan on demand service’. This works out fine and nowadays we produce 15.000 scans a week. When a request for scans is received it takes three to five weeks to produce the scans and bring them online. A lot better than the calculated 420 years…

However, then came the questions from customers that led to VeleHanden. This time the questions were not as simple as ‘where is the button’. They were more like vague complaints: ‘I cannot find it’, ‘It is too difficult’ or ‘Why is there so much’. It took us some time before we comprehended what was needed. Eventually, we came to the conclusion that our customers do want to find things, but they do not want to have to search for them. They want to know the history of their house, or the history of their family, but they do not want to make a study of how the population was administrated in the past. They just want to type in a name or an address. In short, they want an index to all those documents.

We had some indexes made in India, but even in low-wage-countries, this is rather expensive. The average costs of indexing a scan vary from three to ten times the price of producing a scan. It would be absolutely impossible for us to find the money to pay for indexing on a large scale. What we needed for indexing were volunteers, and not just a few, but lots and lots of volunteers. And that is how the idea for this crowdsourcing project was born.

Can we build it?

To attract a real ‘crowd’ to our project it seemed better to us, not to operate on a local scale. A project where all archives in the Netherlands could benefit from, seemed more appropriate. So in 2011 we invited our colleagues to join us in a nationwide pilot project to index all Dutch Militia registers. As enrolment for the army was obligatory, the name of every Dutch 19-year-old boy is written at least once in these registers. So nowadays every man or woman with Dutch ancestors can find some of them in these registers. An interesting project for a real crowd.

However, a nationwide project costs a lot of money. So the first problem to solve was the funding problem. The solution for this problem was found in a public private partnership. You could call it co-creation, as there were various partners involved in making VeleHanden possible.

The first category of partners are the archival institutions. They were asked to pay for the scanning of their own documents. The Amsterdam City Archives offered to manage the scanning process and we were able to negotiate a very low price per page, simply because of the size of the project. So we as archival institutions gain a lot as we receive high quality and cheap scans plus an index and all we learn from being part of an innovative project.

Secondly, VeleHanden was supported by public funds that want to stimulate people to participate in cultural and socially relevant projects. Two public funds decided to join us: the Mondriaan Foundation and VSBfonds.

The third party that profits from the project are the users of the Militia registers website who like to have a look at, or to download our scans. These users pay € 0,50 per scan. Of course there is an important exception: our crowdsourcers can earn scans for free.

And then there was a very important fourth partner in this project: the company that would own and exploit the crowdsourcing platform. We send out a Request for Proposal in which we offered a fixed price for the company that came up with the best plans. We did not ask for lowest price, but we asked for best quality and best warranties for continuity in the future.

We received seven enthusiastic proposals and we decided to contract two partners: one for the crowdsourcing platform VeleHanden.nl and one for the search system and webshop Militieregisters.nl. Geneabase built and owns militieregisters.nl and Picturae exploits the VeleHanden website. Currently, the crowdsourcing platform facilitates already ten institutions in eleven different projects. When an archive service wants to use VeleHanden, it has to pay a service fee that is related to the size, complexity and duration of the project. If new functionality is required for a particular project, it is up to Picturae to develop this functionality in dialogue with the archival institution. However, the archive service retains control over both the digital images and any metadata created by the volunteers during the project. We believe that the partnership therefore combines a commercial imperative for Picturae to support, develop and sustain VeleHanden with the archival institutions’ mission to promote online access and public engagement with archives.

In the next step of development, the actual creation of the platform for crowdsourcing, we invited a lot of people to help us ‘co-create’ VeleHanden. We realised that we needed to attract a huge and diverse community for crowdsourcing. This meant we had to attract a lot of people that we had never met before. Who are they, why would they participate and what is important to them? We needed to get to know these people a bit better. And we felt they had a right to participate in the development of the platform, as they were going to do the actual work on the website.

So we simply asked in one of the City Archives’ newsletters for volunteers to join the project to help us testing the tools for data entry and for quality checks and to test login procedures and forum facilities. In two days’ time 150 people volunteered and we had to stop enrolment for testing as our software-partner was getting a bit nervous.

We started the test period when the first tool, the tool for data entry was ready to be tested. We organised a meeting to get acquainted with our panel and to explain our ideas. They started the data entry, and at the same time we started developing the tool for control. Six weeks later, we held a second meeting, explained the working of the control tool and in the second test period our testers checked the data they had entered in the first six weeks and by doing this they tested the control tool.

All in all the testing took four months. We held another meeting and we organised several surveys, using Google analytics. Attendance to the meetings varied from twenty to fifty people. Even more important than the meetings and the surveys was the Forum. The Forum was a busy meeting-place for direct contact on a daily basis between testers and the project team. And it continues to be; direct communication is a very important feature of the site and we keep thinking of improvement.

All in all, this co-creation of our platform worked out very well and our software engineer told us she wished all projects could be done in this way. She enjoyed explaining how the site worked not only to us, but also to the actual users of the site. And she said it was very rewarding to be able to fix bugs as soon as people see them and to get a nice ‘thank you’ for it, instead of receiving complaints after you thought you had finished your job. On the other side, the volunteers were very happy to receive answers from the software engineer herself and they experienced that we really paid attention to what they said.

During this testing period the core of our crowd was formed. At November 3rd 2011, when we officially baptised VeleHanden and send it out in the world to grow, the testers transformed to ambassadors of the website. They do all the quality checks and they help new arrivals. At this moment, VeleHanden counts more than 2900 members and it still grows daily. So we can assume we managed to create a crowd.

 

VeleHanden.nl

 The platform VeleHanden.nl consists of several parts that we will briefly describe here. Of course this will make even more sense if one would logon to the website and browse around.

On the home page you will find general information about VeleHanden. What is it? Why should I join in in this project? What kind of projects can I work on? Each project has its own page with more specific information about the source the project is about, the task, the reward, relevant news items et cetera. Per project you can also find statistics on the overall progress of the project and which member entered or controlled most scans. We think this is also an important aspect for motivation of the crowd. Most of the projects are still genealogical projects, initiated by archival institutions, but there are also video tagging and photo tagging projects and even one in which you help determining louses and mites. This shows that VeleHanden is still developing. Abb.  Data entry – photo tagging

On the member page (‘Gebruikers’) you find an overview of all volunteers who work on one or more projects. On the Forum people can ask questions or share interesting information with the crowd. The menu items News and Help (FAQ) speak for themselves.

Each member can choose a project to work on, and by clicking on the ‘invoeren’-button, you start with the data entry. Data can be entered once or twice, or even more times, if you would like to. For most genealogical projects double data entry is chosen. A volunteer types the requested data and once he finished the scan and saves the data, a new scan immediately appears. We hope this motivates the volunteers and makes VeleHanden almost addictive. While entering data, the participants can use two buttons to contact the project managers. They can mark a scan as unusable (‘onbruikbaar’), for example when it is a cover of a book, or as remarkable (‘opmerkelijk’), if something special is noted. In the Militia registers project for instance we asked the crowd to use the ‘remarkable-button’ when they recognised famous persons. This led to a nice summing up of heroes on Militieregisters.nl.

 

10_Zeeland_data entry photo VH

 

 

10_Zeeland_data entry text VHAbb 2.  Data entry – archival project

 

 

 

 

The control tool is only accessible if you have been appointed controller by a project manager. The data entered by both participants is shown next to each other and next to the scan and differences are highlighted. The controller can see in an instant where he should pay extra attention to. He can choose one of the entries by clicking on it, or if he disagrees with both, he can enter the correct data himself. If he does not know what is right either, he can use the problem button to send an email to an expert, someone from the archival institution. This option is not used by the controllers often though. As a default setting both participants are rewarded with VeleHanden points, but if the controller suspects misuse, he has the power to take points. This is also a rare exception. Abb. Control tool

 

Pleasure, praise, profit

We saw that we have already passed the problem of making a crowd. Now the important challenge is to maintain a crowd. Of course we have thought a lot about it: it started during the development of VeleHanden, but thinking of ways to keep the crowd still continues. We believe the work has to provide three things to the crowd: pleasure, praise and profit.

First of all: it must be a real pleasure to use the website. It must be user friendly and really fast. The task itself must be fun and it must be nice to come into contact with like-minded people that you get to know after a while.

Then you must feel and be told that the work you are doing is appreciated very much, not only by the archives, but by all historical researchers now and in the future. Working on VeleHanden is a socially relevant activity. It may sound obvious, but this is an important feature of all volunteer work: volunteers have to feel appreciated and praise is a simple, but an important reward. Important at this point is also to show the results: as Amsterdam City Archives we weekly add the newly indexed names in a work in progress index on our website. This is great for our crowd, because if a participant enters a name today, he will be able to search and find that name next week. For us as project team it is great as well, because we see our indexes grow with about 15.000 to 20.000 names a week, which is incredible.

And, last but not least, you have to gain something as well. At VeleHanden you receive what we call VeleHanden points, to be compared to Airmiles. These points are not worth a lot; an estimate would be that you earn € 0,50 an hour. But still, it is a sign of appreciation and we learned this is important to our crowd. Each archive can decide how many points they want to reward their volunteers and can decide what kind of ‘products’ they offer. Usually, it is access to scans, which are otherwise only downloadable for money. But we also have sent flowers or chocolates and organised workshops, guided tours and even a lottery. Although it is not the decisive factor to participate, it is important to reward your crowd. An example that illustrates this, is the birthday of the VeleHanden project of the population registers. We decided that on that particular day, people would be awarded five instead of the regular three points for each scan. Where the average of indexed scans was 250 a day, on the birthday it suddenly rose to 591.

These three p’s, pleasure, praise and profit, are the key factors which we will continue to monitor and develop to maintain our crowd, in hope we can keep them as enthusiastic and hard-working as they are now.

 Results

To conclude some more details about the results of crowdsourcing. In almost two years, VeleHanden counts 2900 participants, mainly from the Netherlands, but also from abroad, for example from Brasil, the United States, Sweden, Australia and even Senegal.

Their production is overwhelming. If I take only the genealogical projects into account, there are over 4,5 million names transcribed. Calculated in working hours, this would take us, regular archivists, more than fifty years. And then we would not even take breaks. So that is huge. On top of that, they really seem to like it, because they are also really active on the Forum. In two years they started 1280 forum topics and responded to them more than 5000 times.

What you see in all crowdsourcing-projects: most of the work is done by a small hard core and this is also the case with VeleHanden. Fifteen participants have processed more than an amazingly amount of 10.000 scans each.

But quantity is not as important as quality. The question most asked is: is crowdsourcing prone to error? Last year, Ellen Fleurbaay and PhD researcher Alexandra Eveleigh from University College London, tackled this question in a paper delivered at the ICA conference. They gave a threefold answer to this question, which we will summarise here.

Firstly, support and guidance is essential for crowdsourcing. Instructions are provided in several different places and formats: as a manual, while indexing and in FAQ. People can ask for help on the Forum or use the remarkable button to bring a scan under extra attention of a controller or project leader. These guidelines and communication tools make the crowd confident.

Secondly, VeleHanden uses a double entry system: as we explained earlier two different people independently index the same scan. And then on top of that, a third person checks the data that the first two have entered. This third volunteer can see both the scanned document and the two sets of data that have already been entered. The third person’s job is to check and decide on the right choice. So three independent persons review a scan, something that would be far too costly if you were not to work with a crowd.

And lastly, a question in return is: what do we consider an error? Sometimes an entry in our index looks like a mistake, but eventually turns out to be right. We will show you one example from the Militia registers.

Searching on the impossible year or birth 1, gives 33 results. So you would think these are 33 mistakes. But take for example Hendrikus Ebeling, who cannot have been born on June 12th in the year 1. As he is registered in 1861, and all boys had to sign on at the age of 19, it is obvious that Hendrikus must have been born in 1842, as were all the other boys registered on the same page. But the strict instruction for the volunteers is to copy exactly what they read in the scan and as you can see, according to the scan his date of birth is June 12th in the year 1. So in the index you can find him born in the year 1 and this is not a mistake made during indexing, but an error in the nineteenth century administration.

So even if you find strange years or dates of birth, you cannot be sure they are mistakes. In the Militia register index of over one million records, there are 316 of these ‘impossible’ years of birth, and this is only 0.027% of the index, a reassuring percentage. Especially if you consider that some of those 316 might even be indexed right.

So all in all we are very satisfied with the development and results of VeleHanden so far. Of course, there are still points of improvement and we still learn every day.

Nelleke van Zeeland
Stadsarchief Amsterdam

nvanzeeland@stadsarchief.amsterdam.nl

 

Ellen Fleurbaay

Stadsarchief Amsterdam

heft 80 - 01-2014 - seiten001-068.inddDieser Beitrag ist ebenfalls inder aktuellen Archivpflege, Heft 80 abgedruckt.

Quelle: http://archivamt.hypotheses.org/759

Weiterlesen