1. As complexity and number of required metadata fields increase, the work of maintaining them rapidly multiplies. As well, accuracy and correction of metadata requires frequent attention. Physical libraries are often bemoaned as on their way out, but how vital will the information professions be to maintaining accessibility of digital data?
2. JISC recommends using an existing ‘standard’, which can be broadly defined as “commonly used and consistently applied formats or processes, which are measurable, well documented, and endorsed by someone.” Even narrowing the definition by that much, there are still a plethora of standards available. How easily can one navigate the many options available?
3. JISC Media appears to be a UK-based service based around assisting educational institutions with digitization, including paid consultancy services. Is the metadata- and digitization field one that is complex and fast-growing enough to be both a prime source on metadata advice and a sector that can support consultation employment?
1. When conducting the VADA project with Dr. Trace, we confronted the disadvantages of community-based metadata. We are trying to mine all the level and collection size information from all the find aids (metadata) in Taro. These finding aids are validated against ead.dtd, and work well in the archival field. But there are too many variations in the xml value fields and a lot of human typos, which cause the finding aids very hard to parse by program.
2. When talking about where metadata is kept, the author lists (a) within the digital file; (b) within a database; (c) in a separate XML-encoded file; or (d) all of the above at once. But what about JSON? JSON promises the same benefits of interoperability and openness as XML, but is more suited to data-exchange and match the data model of most programming language. So is JSON also adopted in keeping the metadata?
1. While I'm all for having variety when it comes to standards, because organization is going to have different needs, given the importance of metadata both currently and going forward, wouldn't it make more sense to pick one and stick with it by now? Instead of having 4-5 different schemas to work with, why not make one and create different sets of rules for different institutions? I think my question about this stems from the fact libraries have used MARC for so long, altering it when and where they need to, but remaining loyal to it for the most part.
2. In all the talks about crowd sourcing, I often get the impression when people say archivists would need to make sure the metadata from users was 'good' feels like a rushed judgment. For instance NASA will release photos taken by the telescopes to both professional and 'couch' astronomers to help them map out celestial bodies seen in each photo. If we wanted a higher chance of good crowd-sourced metadata... wouldn't the logical conclusion be to ask experts in the field of materials we're digitizing to help out with it, rather than unleashing it on the Internet at large?
1 - While MARC provides a template and consistent field labels, it's not an actual content standard: MARC cataloguing uses either AACR2 or RDA, which are standardized versions of how the fields should be populated. I think one of the primary challenges of picking one metadata standard is the variety of materials that are being described, hence the Shoah Foundation developing their own metadata to describe the testimonies in the archive. Even AACR2 and RDA have come under fire because the LOC subject headings are still rather slow to change or be updated when it comes to language, in particular language dealing with racial identities, sexual/gender identity, and language surrounding things like sex work or marginalized people/professions. It would be great to come up with a comprehensive standard, but realistically the needs of institutions are so varied, and as communities become more involved in the description of their own materials, it will become even more challenging to pick a single standard.
2. I think one benefit of crowdsourcing is lower cost, which we cannot expect much from the experts. And non-professional would also generate high quality metadata, I believe, which could be improved by a better interface and a better mechanism for the crowdsourcing task. And maybe the amateurs don't know much about processing archival materials, and their metadata may not fit the format of the archival discipline. But there are still values from the data, from which we can extract the correct format and valuable information (the progression from data to information per se).
1 - Given that one of the biggest metadata related choices a repository has to make is now many "layers" of data to include, I would like to know if anyone has resources or studies that indicate the kinds of metadata users need to gain access to relevant materials. Do people generally search by the title of an item, the author, do they free-form browse under subject headings? Does descriptive metadata deal with most of the users' needs, or are they also using administrative data (e.g., looking for a specific filename in a series) to locate related objects?
2 - Having spent a lot of time entering repetitive metadata for digital objects in another class, I'm curious if there are useful ways one can automate the generation of metadata that could be overseen or managed by community members. I'm not 100% discussing crowdsourcing, more the idea that automated metadata generation could be overseen by a volunteer/intern, who would then do quality control/cleanup. Given how much the Shoah foundation spends on each individual interview, it seems like an automated system would be more cost-effective in the long term, and possibly account for user inconsistency in generating metadata.
2 - This is an interesting idea, and one I've been surprised hasn't shown up more in the readings. There seems to be a general attitude that one must choose between full unsupervised automation and wholly manual metadata / transcription / cataloging / etc., which seems unrealistic given the obvious limitations of automation. While we are working to improve the technology, it makes sense in the meantime to teach the technology to recognize when it needs help from a live person. This would maximize efficiency by capitalizing on the strengths of both humans and automated technologies.
1. The author in this guide discussed the different kinds of metadata and the terms used to describe a piece (creator, photographer, etc…). The example used is the Mona Lisa, painted by Leonardo da Vinci, and the author names da Vinci as “creator”. However, I know that, when cataloging, da Vinci would have been named as “artist”. Should these kinds of relator terms be more standardized? Would this kind of variation (one person using “creator” while another “artist”) prove confusing in the long run? Does it matter?
2. The author talks about where metadata is created, and claims that many times metadata comes from libraries, museums, etc., and that, while this can be a good thing, it can mean that these institutions “can also incorporate old-fashioned or 'legacy' approaches that may have worked well in a non-digital environment, but are not as practical or useful in the digital world”. Are libraries and museums really so inflexible that they wouldn’t keep up with what users need? Is the solution to have more user involvement? The author mentioned that libraries and museums generally have a pretty high standard for metadata – is it worth possibly sacrificing these standards by handing more responsibility to users?
3. The author discusses the different communities that use metadata, and how the vocabularies among these communities differ. For example, the relator term “creator” may work for one community but not another. However, as the author points out, sometimes multiple communities are looking at the metadata for a work or are using the same databases. The author’s solution is to “take a more generic approach [to the metadata], which often means some kind of compromise”. What might this comprise look like? Is it possible to use different forms of metadata in the same database on the same record? Would that just confuse things?
1. This is why we have metadata/cataloging standards like AACR2. What I think you're getting at is that, across disciplines, this could certainly cause some issues. But I don't think it would be feasible to have consistent standards across all disciplines and for every situation.
1 It is said in the paper that metadata usually has two part: metadata schema categories and metadata vocabulary terms. When we retrieve the metadata, which part do we use, or we can use both categories from schema and vocabulary terms?
2 There have different purposes of metadata. And the author also mentions the issue of overlapping. I have a question about that when we describe certain item, do we describe all factors for different purposes or just for the certain purpose? If it is the latter, how about the overlapping of the work human devote into to describe it?
3 Who are the users of metadata? As it is said in the paper, implicit metadata can be "users" of other metadata from digital resource. And some users are human beings. So is there's a different editing different kind of metadata for the same object?
1. In the article, the author states "(metadata) makes it easier to retrieve the image in a search, since the search query can be much more specific, targeting relevant categories rather than searching across all of the metadata...". Since I've worked in an image search engine I know this is not the truth. Tagging metadata for a small number of images might be a solution, it's impossible to tag every image on the web, not to mention tagging cannot cover every aspect of an image. Google Image Search once tried tagging images by man yet turned out a failure. In this sense, do metadata really have an advantage over traditional ways of retrieving images?
2. Concerning metadata, one thing I am interested in is how to build metadata for videos. Videos are more content-rich and complicated than articles or images. On the Youtube, tags with a videos mostly are abstract and vague. Then, is it possible or how to build a metadata framework for videos on the web?
3. By the end of the article, the author introduced XML as a typical format of using metadata. I'm wondering how the future of XML would be with the development of web technologies as well as metadata itself. Would metadata serve for more entities in daily lives as it serves the entities on the web?
1. Under the section titled ‘Different Levels and Layers’, they show an example of the Mona Lisa and the different types of metadata created for different versions of the object. Is all of this metadata ultimately useful, or do users get lost in a metadata soup? Is more and more metadata a good thing – creating more information – if the standards vary so much between institutions? For example, would the metadata on the Mona Lisa from the Louvre vary from some institution that was using a digital replica of the Mona Lisa for some project or other?
2. I wonder if most metadata really is user-friendly or if it’s only useful to those who understand why we need it and how to create it. Does is just depend on the type of metadata created that impacts the ultimate user and defines utility?
3. Is the whole metadata trend and use just a re-branding of something else? Similar like what we have discussed about the conversion of library schools to iSchools? As a system for classification, is it really more useful or comprehensive than what came before it? Is the advent of metadata just a response to the new digital age where traditional cataloging and classification couldn’t quite follow?
1. When talking about selective, this article says ‘The challenge for those applying metadata to a digital collection is to work out which information is going to be the most important and useful to record’. Why is this a challenge? In practice, how could we decide which information is the most important? Are there any standards? What can we do with metadata?
2. ‘Rather than randomly associating terms with the digital file, it is common to use a set of generic categories.’ Why do we commonly use this categories? I think there must be some limitations when we use this kind of categories, since not all the digital files can be described by the same ‘set of generic categories (e.g. 'Creator', 'Title', 'Subject')’.
3. ‘In developing a metadata framework for your collection it is important that you identify all of these users and needs. It's best to ask your users what information they need rather than make assumptions.’ Is that true? Do people really do user research when they create metadata? If so, what do they do?
1. I really enjoyed reading this introduction in conjunction with the Metacrap website by Doctorow. If humans are inherently bad at applying consistently good metadata to documents, do technologies exist that could assist us in attributing consistently accurate metadata during file creation? I am thinking of googles current project that allows you to search the web with by taking a picture of an object. If this data can accurately predict what objects are in a picture for example, could we apply this metadata to pictures throughout Flickr? I know that similar types of things already exist in iPhoto with face recognition but I would be really interested to see what else is out there.
2. I am curious how one comes about deciding upon a specific taxonomy for descriptive metadata of images in a setting like a library? It would seem that certain images would have a mass appeal to different audience who all might be ascribing different search terms based on their fields. How does one know when you have been descriptive enough? Can you be too descriptive?
3. I am curious how archivists deal with a creator’s metadata after death. During my time archiving photographs for an individual I would create fonds of images based on subject matter such as dogs and trees. I can see how certain images might be of interest to people under different contexts and I am wondering how an archivist would deal with this problem while still retaining an accurate record of how the author had chosen to organize the work.
1. We see that there are different purposeful metadata: resource discovery, descriptive, provenance and rights, technical, administrative, preservation, and structural. Are there best practices for knowledge management systems that might use a combination of these purposeful metadata schemas?
2. I talked a little about this when discussing archival information - can software auto-embed metadata and also make user generated metadata required into the document? For example, most businesses use MS Office products. If archivists and other IS professionals worked with Microsoft to develop a minimum required set and then maybe industry-specific required metadata, it may help to alleviate businesses burdened with many files that don't have any metadata attached. Businesses are left with millions of files that they are not able to retrieve during a search or even know that the document exists.
3. This week we have read articles that discussed some of the problems encountered when using metadata. But the recent stories (such as those concerning Snowden leaking how the government is tracking your activities) present a different story about metadata. It gives the impression of being incredibly precise. Is there a difference between the metadata being collected electronically through devices now and the metadata provided to Google Books (say from the Harvard Library)? What about all of the different metadata types - are these devices all using the same type?
1) Are controlled vocabularies just as important as they ever were? With more and more libraries and archives relying on crowd sourcing in order to come up with more subjective and varied search terms, might controlled vocabularies stand in our way?
2) How subjective is implicit or intrinsic metadata? Though this data is supposed to be cut and dry, and unchanging in nature can we still bend it?
3) Where should metadata be kept? How do we make sure this invaluable information isn't lost somehow?
1. Since different metadata standards use different approaches to describe different levels of a digital resource, I wonder will people use different standards when they build a website together? If so, they might describe a same level of digital resource in different ways, will they cause a big mess? I think there must be a way to support people use different standards together. If so, how does it work?
2. The author said it is important to identify all of these users and needs when developing a metadata framework. However, I found it could be very difficult to take every user and user’s need into account. If it is a big project, how can we make sure that we identify all users and their needs?
3. The author mentioned that it might be possible to get your collection users to add to the metadata in a semi-controlled way. I wonder what is the semi-controlled way, is there an uncontrolled way for users to add metadata? Where can we keep the metadata from end users? How will we evaluate the validity of these tags and annotations from users?
Is there a little bit of a disconnect in defining the metadata categories as “schema” and having the way of encoding metadata categories fall under the same name while at the same time having specific terminology for metadata being preferred? It seems to be a situation of do as I say not as I do in this situation.
With the inclusion of layers and even sub-categories for metadata, is there a point at which there is too much metadata or does having all of that metadata act as a benefit for both the users and archivists/catalogers/creators?
With the advent of social media, there exists a seemingly infinite amount of use metadata especially in the form of tagging. How might we use the other types of metadata to work through the noise of tags to discover the “real or useful” metadata or would that be disruptive in and of itself?
1. Metadata can have various definitions like it can be used to ‘tag’ an image, can be the file size of the digital file or something more complex like the subject matter of the resource or legal rights associated with its use. Considering these examples, is metadata the shredded form of the original data? If it contains the subject matter, then how can metadata be a descriptor to the original data? 2. Various kinds of Schema are discussed such as the DC Schema, SEPIADES schema, the METS schema etc. Which of these is more efficient than the other and why? Can we use one of these or form universal schema that can be used in all databases? 3. The two types of metadata are implicit and explicit metadata. Implicit metadata is more useful to the administrators than the users. Is there a way to delete them when they are redundant or not used? Will this reduce the value of the data ? Can the system become more efficient by doing this as it frees up memory space?
1. Is metadata about metadata categorized/structured in the same way as traditional metadata? 2. In consideration of intrinsic and extrinsic metadata -- is there ever an instance where metadata is both? If so, what does that look like, and how would it be classified? 3. I'm curious as to how metadata changes. For example, if information about an object is modified over time, is the original metadata forever associated with it? Are there formal processes or frameworks for addressing changing metadata information?
1. At one point, the author writes that “metadata invariably offers a selective or simplified description of a source”. Does this imply that metadata is often subjective? Is the inclusion of certain metadata dependent upon the repository?
2. The author discusses how metadata is often held within databases but can, if decided, be embedded within the digital file itself. Can it be argued that one or other of these approaches is better or more effective? Doesn’t embedding within the file cause issues in that then the file changes every time you add new metadata, changing its original form, whereas storing it in a database allows the file to stay static while metadata can be added at any point and any time?
3. In discussing standards, the author differentiates standards base on administrative versus descriptive. They go on to say that “there have also been some attempts to create standards that are focused on particular purposes”. This seems like a very vague statement. Is it written as such because delving deeper is not within the scope of the paper, or if not, what exactly are these “particular purposes”? He cites PREMIS, which I’m more or less unfamiliar with.
1. I think that metadata is subjective. For archives and libraries,a repository decides the minimum number of fields that must be added to a catalog record. Additional metadata added could be added based purely on what the cataloger feels is important to add, and this can be impact by their own experiences.
I am also intrigued by metadata that is added by different users based on the different purposes of the object in its life cycle. For example, a person may take a photo of their mother and label it "Mom." Many years later, this photo could end up in a historical society, and it may be kept for historical reasons because the mom is standing next to a historical building. A cataloger there may re-title the photograph based on the name of the building, no longer really caring who the woman is or they might not even know. Here I think the metadata is subjective based on the reason the object is deemed important by different people. I am wondering if there is a way to embed in the object all of the metadata changes the object may have applied to it in its lifecycle.
1) Like many of our readings, this one called to mind some of the difficulties of language-based information description and retrieval when handling non-language-based resources. Some descriptions of their example, the Mona Lisa, are obvious (e.g. “woman” and “Renaissance”), but other potentially important or interesting information about the work may be less obvious, like the fact that it was commissioned by a wealthy man as a portrait of his wife, or the presence of a landscape in the background. How do metadata creators draw the line between vital information and the merely potentially useful? Does it vary by collection?
2) The article points out some of the difficulties caused by “legacy” metadata schemas, in fields where pre-digital information retrieval systems have led to a schema that has since become entrenched even as its usefulness has waned. How is it possible to modernize metadata schemas while still accounting for the needs and expectations of the established community? Is it better to compromise and gradually evolve, to implement a new and unfamiliar schema that suits the information’s needs better, or to adhere completely to the possibly-outdated schema?
!. If cost was not an issue in the creation of descriptive metadata, would more necessarily be better? In looking at the Mona Lisa example, it seems that a very simplified approach is sufficient for a single information object and that too much descriptive metadata would make retrieval more difficult. What types of objects would benefit from a lot of descriptive metadata?
2. The author compares HTML to XML saying that HTML is aimed at presentation while XML focuses more on meaning. The example given doesn't do much in the way of clarification. What does the author mean by this?
3. When discussing implicit metadata, the author states that at times some is extracted and stored elsewhere. The reasons he gives seem very valid but he says that it is typically not done. Why wouldn't this be done? Is it because of cost and time issues?
1. I have never heard of “resource discovery” metadata before this article. The article mentions that it is the metadata that helps us find a resource. I am wondering if this is a particular metadata field, or if this is more of a theoretical concept that can change with each search. For example, I could search for a book in a library catalog by its title. Another person may search by the author’s name. Does this make the title or the author the resource discovery metadata in each case, or are both in total the book’s resource discovery metadata?
2. The author mentions that sometimes particular communities will “incorporate old-fashioned or ‘legacy’ approaches that may have worked well in a non digital environment, but are not as practical or useful in the digital world.” I am wondering what examples of legacy metadata exists in the archives and libraries fields. Are there aspects of archival metadata, such as a specific field or format of a finding aid, still included on digital finding aids, merely because they were documented in a certain way on paper, but might make sense to represent a different way now that the finding aid is digital? Were there metadata practices that were carried over into MARC from card catalogs, that doesn’t make the best sense given the digital aspect of MARC records?
1. " Metadata is usually structured in some way." Are there Metadata standards or uses that don't have any structure and just freely associate metadata with the object? It seems to me that doing this would not be very productive. Sure, the information is still stored and associated with the object but if it's not in some structured format I don't see how it could be useful.
2. "As with the different levels, metadata schemas tackle this problem in different ways. Some will create sub-categories (e.g. Creator_OriginalWork; Creator_SurrogateImage); others, separate categories (e.g. Artist; Photographer; Scanning Technician); others, completely separate records for each layer." What is the advantage of keeping separate records for each layer? If someone is interested in a scanned image of the Mona Lisa I may not want t know who the scanning technician is but I would rather have the information and not need it in the case that I may want to know later down the road.
3. What are the advantages of storing metadata in a closed system like a database or within the actual file itself? I would think that storing metadata in an open and accepted format like XML so that if he institution that is curating these files has the most options available to them.
1. Under the section, “What is Metadata?” it says that metadata is often held within databases, but it can take other forms - it can just as easily be found embedded within the digital file itself." Does metadata take other forms other than being “data about data”? Also, does metadata only exist in print or online records? Could a person be metadata somehow? Ex.) Say a person is wearing a shirt to advertise something. Is that person essentially data showing where to find other data or resources?
2. At what point does metadata end? Is it when it gets updated or becomes obsolete? Or is there metadata about when other metadata was created or destroyed and so on ad infinitum?
3. In talking about intrinsic and extrinsic metadata it says that extrinsic data is created by humans whereas intrinsic metadata is technical in nature. Since technology is created by humans doesn't that make intrinsic metadata extrinsic in nature as well?
1. The article briefly mentions that intrinsic metadata is recorded or collected with little or no human intervention, like a digital photo retaining information about the type of camera that took the picture. I remember this being an issue in the news recently, with photos taken with iPhone cameras and uploaded to Facebook/the Web making this metadata, including time and location that the picture was taken, available to tech-savvy users. The average non-iSchooler iPhone user probably wouldn't know that this information is being collected, but it can prove to be a huge security risk. How big a matter of professional ethics should it be for this information about metadata practices (meta-metadata, if you will) to be publicized about products?
2. How possible might it one day be to make extrinsic metadata automatically available, without the high level of human involvement currently required? For instance, with OCR and other document reading technologies, would it be possible to write a program that reads documents in a collection, seeks particular collocations or patterns, and generates descriptive metadata about that document? (This may already exist; forgive my oblivion if that's the case.)
1. Metadata might be simply a free text description or set of keyword used to annotate or ‘tag’ an image. It is sure free text will mike metadata easy to generate, but will it cause the metadata hard to be understand or retrievable?
2. Metadata aims to cover so many functions, like what it is, where the resource has come from, who owns it… But is it efficient? A metadata that might beyond the size of original data will increase the workload of database.
3. Metadata should serve communities and users, so it can take many different forms. The different standards will make metadata not compactible in other system. So should we create a universal standard for the metadata, or part of it?
1. Are pre-digital forms of descriptive information metadata? Are encyclopedias metadata? Dictionaries? Graffiti tagging? Deeds and titles? The Domesday Book?
2. How has crowd-sourced/open metadata creation influenced professional library, archive, and museum practice? How has it affected academic theory on metadata? Will it be possible for information professionals to retain authority over metadata in an increasingly open-access, online information universe? How must we reconceptualize our roles and responsibilities as metadata authorities?
3. What, if any, programming language and/or software is being developed to "merge" various metadata standards into a larger, multi-coded metadata system (e.g. Dublin Core and METS)? What are the possibilities and pros and cons of such an impulse? Will it ever be possible - or desirable - to have one, all-encompassing, multi-multi-field meta-metadata schema?
1. It seems like metadata could be divided continuously as long as the data managers wish. For example, the location of "Mona Lisa" can be also divided into museum, city, state, and country.Some patrons might want to get to know these information, while others might find it useless. So,which aspects might be taken into consideration when we want to decide how detailed the metadata should be?
2. The author claims that there are four categories or types of metadata: descriptive metadata, administrative metadata, structural metadata, and use metadata. Is it necessary and possible to collect all four types metadata for one document?
3. In the "Different communities and users" part, the author says that "We've just distinguished the end users of the collection from those managing the collection ( who are also "users")." What differences do these two kinds of users have when they are using metadata?
1. In this article the author states that there are different schemas of metadata that are created for specific purposes. These types of metadata include descriptive, administrative, structural, and use metadata. Since there are different schemas that focus on these different types of metadata then is it possible to have two different schemas of metadata that are applied to the same collection? If so then what are the benefits and drawbacks of doing this? 2. In this article the author states that use metadata is metadata that is created by or about the users. We have seen that crowdsourcing is a viable method for getting users to create metadata to describe the contents of a collection. Would you consider this to be use metadata or descriptive metadata since it is useful to describe the collection? 3. In this article the author states that there are three different places that metadata about an object can be found. It can be found inside the information object, in a separate XML file, or in a database. Which of these three places do you think is the best place to store metadata and why? Which of these places do you think is a bad place to store metadata and why?
1) If an advantage of metadata is to share the metadata with other similar collections, how does choosing to describe items according to different layers affect the share-ability of the metadata? If one institution wants to create sub-categories, another separate categories, or another separate records for each layer are these records useful to other institutions or users trying to explore the collection? 2) In the section “Where does metadata come from?” it is stated that there is intrinsic and extrinsic metadata, the difference being whether a human or machine created it. Isn't automatic metadata ultimately also created and associated with a resource by the humans who created the programs that automate the metadata? This was simple and obvious metadata based on file characteristic, formats, and origins of the file, but is there an essential role for information professionals to be involved in the design of programs that produce automatic metadata and to what level should they be involved? 3) Metadata can be stored in different places depending on what it is and what it is used for, some stored with the file other just through association. What are the advantages to combining or separating administrative, descriptive, structural, preservation, and use metadata and does it depend on the resource you are describing, a random photograph vs. the Mona Lisa?
Q1. The article mentioned the image metadata and set the portrait Mona Lisa as an example to illustrate that metadata is usually structured in some way. I am wondering who is responsible for selecting the metadata elements that associated with an image? When factors do they need to consider when making the decisions? Q2. In the Different purposes section, the author states that metadata can serve different purposes, like help us find resource or tell us what it is. So I want to know which purpose is the one that promote the emergence of metadata? Q3. The author talked about the metadata applied in different communities and facing different users. Is there a role who is exclusively in charge of metadata management in any kinds of communities? How do they tackle the outdated metadata?
1. I can imagine where one comes across “resource discovery” metadata, or “descriptive" metadata but where does one come across “provenance and rights” or “technical” metadata?
2. It seems like the more one adapts a metadata schema to a particular user community, the less potential it has to become more universally searchable as it won’t be compatible with any other metadata schemata. Isn’t this problematic?
1. As complexity and number of required metadata fields increase, the work of maintaining them rapidly multiplies. As well, accuracy and correction of metadata requires frequent attention. Physical libraries are often bemoaned as on their way out, but how vital will the information professions be to maintaining accessibility of digital data?
ReplyDelete2. JISC recommends using an existing ‘standard’, which can be broadly defined as “commonly used and consistently applied formats or processes, which are measurable, well documented, and endorsed by someone.” Even narrowing the definition by that much, there are still a plethora of standards available. How easily can one navigate the many options available?
3. JISC Media appears to be a UK-based service based around assisting educational institutions with digitization, including paid consultancy services. Is the metadata- and digitization field one that is complex and fast-growing enough to be both a prime source on metadata advice and a sector that can support consultation employment?
1. When conducting the VADA project with Dr. Trace, we confronted the disadvantages of community-based metadata. We are trying to mine all the level and collection size information from all the find aids (metadata) in Taro. These finding aids are validated against ead.dtd, and work well in the archival field. But there are too many variations in the xml value fields and a lot of human typos, which cause the finding aids very hard to parse by program.
ReplyDelete2. When talking about where metadata is kept, the author lists (a) within the digital file; (b) within a database; (c) in a separate XML-encoded file; or (d) all of the above at once. But what about JSON? JSON promises the same benefits of interoperability and openness as XML, but is more suited to data-exchange and match the data model of most programming language. So is JSON also adopted in keeping the metadata?
1. While I'm all for having variety when it comes to standards, because organization is going to have different needs, given the importance of metadata both currently and going forward, wouldn't it make more sense to pick one and stick with it by now? Instead of having 4-5 different schemas to work with, why not make one and create different sets of rules for different institutions? I think my question about this stems from the fact libraries have used MARC for so long, altering it when and where they need to, but remaining loyal to it for the most part.
ReplyDelete2. In all the talks about crowd sourcing, I often get the impression when people say archivists would need to make sure the metadata from users was 'good' feels like a rushed judgment. For instance NASA will release photos taken by the telescopes to both professional and 'couch' astronomers to help them map out celestial bodies seen in each photo. If we wanted a higher chance of good crowd-sourced metadata... wouldn't the logical conclusion be to ask experts in the field of materials we're digitizing to help out with it, rather than unleashing it on the Internet at large?
1 - While MARC provides a template and consistent field labels, it's not an actual content standard: MARC cataloguing uses either AACR2 or RDA, which are standardized versions of how the fields should be populated. I think one of the primary challenges of picking one metadata standard is the variety of materials that are being described, hence the Shoah Foundation developing their own metadata to describe the testimonies in the archive. Even AACR2 and RDA have come under fire because the LOC subject headings are still rather slow to change or be updated when it comes to language, in particular language dealing with racial identities, sexual/gender identity, and language surrounding things like sex work or marginalized people/professions. It would be great to come up with a comprehensive standard, but realistically the needs of institutions are so varied, and as communities become more involved in the description of their own materials, it will become even more challenging to pick a single standard.
Delete2. I think one benefit of crowdsourcing is lower cost, which we cannot expect much from the experts. And non-professional would also generate high quality metadata, I believe, which could be improved by a better interface and a better mechanism for the crowdsourcing task. And maybe the amateurs don't know much about processing archival materials, and their metadata may not fit the format of the archival discipline. But there are still values from the data, from which we can extract the correct format and valuable information (the progression from data to information per se).
Delete1 - Given that one of the biggest metadata related choices a repository has to make is now many "layers" of data to include, I would like to know if anyone has resources or studies that indicate the kinds of metadata users need to gain access to relevant materials. Do people generally search by the title of an item, the author, do they free-form browse under subject headings? Does descriptive metadata deal with most of the users' needs, or are they also using administrative data (e.g., looking for a specific filename in a series) to locate related objects?
ReplyDelete2 - Having spent a lot of time entering repetitive metadata for digital objects in another class, I'm curious if there are useful ways one can automate the generation of metadata that could be overseen or managed by community members. I'm not 100% discussing crowdsourcing, more the idea that automated metadata generation could be overseen by a volunteer/intern, who would then do quality control/cleanup. Given how much the Shoah foundation spends on each individual interview, it seems like an automated system would be more cost-effective in the long term, and possibly account for user inconsistency in generating metadata.
2 - This is an interesting idea, and one I've been surprised hasn't shown up more in the readings. There seems to be a general attitude that one must choose between full unsupervised automation and wholly manual metadata / transcription / cataloging / etc., which seems unrealistic given the obvious limitations of automation. While we are working to improve the technology, it makes sense in the meantime to teach the technology to recognize when it needs help from a live person. This would maximize efficiency by capitalizing on the strengths of both humans and automated technologies.
Delete1. The author in this guide discussed the different kinds of metadata and the terms used to describe a piece (creator, photographer, etc…). The example used is the Mona Lisa, painted by Leonardo da Vinci, and the author names da Vinci as “creator”. However, I know that, when cataloging, da Vinci would have been named as “artist”. Should these kinds of relator terms be more standardized? Would this kind of variation (one person using “creator” while another “artist”) prove confusing in the long run? Does it matter?
ReplyDelete2. The author talks about where metadata is created, and claims that many times metadata comes from libraries, museums, etc., and that, while this can be a good thing, it can mean that these institutions “can also incorporate old-fashioned or 'legacy' approaches that may have worked well in a non-digital environment, but are not as practical or useful in the digital world”. Are libraries and museums really so inflexible that they wouldn’t keep up with what users need? Is the solution to have more user involvement? The author mentioned that libraries and museums generally have a pretty high standard for metadata – is it worth possibly sacrificing these standards by handing more responsibility to users?
3. The author discusses the different communities that use metadata, and how the vocabularies among these communities differ. For example, the relator term “creator” may work for one community but not another. However, as the author points out, sometimes multiple communities are looking at the metadata for a work or are using the same databases. The author’s solution is to “take a more generic approach [to the metadata], which often means some kind of compromise”. What might this comprise look like? Is it possible to use different forms of metadata in the same database on the same record? Would that just confuse things?
1. This is why we have metadata/cataloging standards like AACR2. What I think you're getting at is that, across disciplines, this could certainly cause some issues. But I don't think it would be feasible to have consistent standards across all disciplines and for every situation.
Delete1 It is said in the paper that metadata usually has two part: metadata schema categories and metadata vocabulary terms. When we retrieve the metadata, which part do we use, or we can use both categories from schema and vocabulary terms?
ReplyDelete2 There have different purposes of metadata. And the author also mentions the issue of overlapping. I have a question about that when we describe certain item, do we describe all factors for different purposes or just for the certain purpose? If it is the latter, how about the overlapping of the work human devote into to describe it?
3 Who are the users of metadata? As it is said in the paper, implicit metadata can be "users" of other metadata from digital resource. And some users are human beings. So is there's a different editing different kind of metadata for the same object?
1. In the article, the author states "(metadata) makes it easier to retrieve the image in a search, since the search query can be much more specific, targeting relevant categories rather than searching across all of the metadata...". Since I've worked in an image search engine I know this is not the truth. Tagging metadata for a small number of images might be a solution, it's impossible to tag every image on the web, not to mention tagging cannot cover every aspect of an image. Google Image Search once tried tagging images by man yet turned out a failure. In this sense, do metadata really have an advantage over traditional ways of retrieving images?
ReplyDelete2. Concerning metadata, one thing I am interested in is how to build metadata for videos. Videos are more content-rich and complicated than articles or images. On the Youtube, tags with a videos mostly are abstract and vague. Then, is it possible or how to build a metadata framework for videos on the web?
3. By the end of the article, the author introduced XML as a typical format of using metadata. I'm wondering how the future of XML would be with the development of web technologies as well as metadata itself. Would metadata serve for more entities in daily lives as it serves the entities on the web?
1. Under the section titled ‘Different Levels and Layers’, they show an example of the Mona Lisa and the different types of metadata created for different versions of the object. Is all of this metadata ultimately useful, or do users get lost in a metadata soup? Is more and more metadata a good thing – creating more information – if the standards vary so much between institutions? For example, would the metadata on the Mona Lisa from the Louvre vary from some institution that was using a digital replica of the Mona Lisa for some project or other?
ReplyDelete2. I wonder if most metadata really is user-friendly or if it’s only useful to those who understand why we need it and how to create it. Does is just depend on the type of metadata created that impacts the ultimate user and defines utility?
3. Is the whole metadata trend and use just a re-branding of something else? Similar like what we have discussed about the conversion of library schools to iSchools? As a system for classification, is it really more useful or comprehensive than what came before it? Is the advent of metadata just a response to the new digital age where traditional cataloging and classification couldn’t quite follow?
1. When talking about selective, this article says ‘The challenge for those applying metadata to a digital collection is to work out which information is going to be the most important and useful to record’. Why is this a challenge? In practice, how could we decide which information is the most important? Are there any standards? What can we do with metadata?
ReplyDelete2. ‘Rather than randomly associating terms with the digital file, it is common to use a set of generic categories.’ Why do we commonly use this categories? I think there must be some limitations when we use this kind of categories, since not all the digital files can be described by the same ‘set of generic categories (e.g. 'Creator', 'Title', 'Subject')’.
3. ‘In developing a metadata framework for your collection it is important that you identify all of these users and needs. It's best to ask your users what information they need rather than make assumptions.’ Is that true? Do people really do user research when they create metadata? If so, what do they do?
1. I really enjoyed reading this introduction in conjunction with the Metacrap website by Doctorow. If humans are inherently bad at applying consistently good metadata to documents, do technologies exist that could assist us in attributing consistently accurate metadata during file creation? I am thinking of googles current project that allows you to search the web with by taking a picture of an object. If this data can accurately predict what objects are in a picture for example, could we apply this metadata to pictures throughout Flickr? I know that similar types of things already exist in iPhoto with face recognition but I would be really interested to see what else is out there.
ReplyDelete2. I am curious how one comes about deciding upon a specific taxonomy for descriptive metadata of images in a setting like a library? It would seem that certain images would have a mass appeal to different audience who all might be ascribing different search terms based on their fields. How does one know when you have been descriptive enough? Can you be too descriptive?
3. I am curious how archivists deal with a creator’s metadata after death. During my time archiving photographs for an individual I would create fonds of images based on subject matter such as dogs and trees. I can see how certain images might be of interest to people under different contexts and I am wondering how an archivist would deal with this problem while still retaining an accurate record of how the author had chosen to organize the work.
1. We see that there are different purposeful metadata: resource discovery, descriptive, provenance and rights, technical, administrative, preservation, and structural. Are there best practices for knowledge management systems that might use a combination of these purposeful metadata schemas?
ReplyDelete2. I talked a little about this when discussing archival information - can software auto-embed metadata and also make user generated metadata required into the document? For example, most businesses use MS Office products. If archivists and other IS professionals worked with Microsoft to develop a minimum required set and then maybe industry-specific required metadata, it may help to alleviate businesses burdened with many files that don't have any metadata attached. Businesses are left with millions of files that they are not able to retrieve during a search or even know that the document exists.
3. This week we have read articles that discussed some of the problems encountered when using metadata. But the recent stories (such as those concerning Snowden leaking how the government is tracking your activities) present a different story about metadata. It gives the impression of being incredibly precise. Is there a difference between the metadata being collected electronically through devices now and the metadata provided to Google Books (say from the Harvard Library)? What about all of the different metadata types - are these devices all using the same type?
1) Are controlled vocabularies just as important as they ever were? With more and more libraries and archives relying on crowd sourcing in order to come up with more subjective and varied search terms, might controlled vocabularies stand in our way?
ReplyDelete2) How subjective is implicit or intrinsic metadata? Though this data is supposed to be cut and dry, and unchanging in nature can we still bend it?
3) Where should metadata be kept? How do we make sure this invaluable information isn't lost somehow?
1. Since different metadata standards use different approaches to describe different levels of a digital resource, I wonder will people use different standards when they build a website together? If so, they might describe a same level of digital resource in different ways, will they cause a big mess? I think there must be a way to support people use different standards together. If so, how does it work?
ReplyDelete2. The author said it is important to identify all of these users and needs when developing a metadata framework. However, I found it could be very difficult to take every user and user’s need into account. If it is a big project, how can we make sure that we identify all users and their needs?
3. The author mentioned that it might be possible to get your collection users to add to the metadata in a semi-controlled way. I wonder what is the semi-controlled way, is there an uncontrolled way for users to add metadata? Where can we keep the metadata from end users? How will we evaluate the validity of these tags and annotations from users?
Is there a little bit of a disconnect in defining the metadata categories as “schema” and having the way of encoding metadata categories fall under the same name while at the same time having specific terminology for metadata being preferred? It seems to be a situation of do as I say not as I do in this situation.
ReplyDeleteWith the inclusion of layers and even sub-categories for metadata, is there a point at which there is too much metadata or does having all of that metadata act as a benefit for both the users and archivists/catalogers/creators?
With the advent of social media, there exists a seemingly infinite amount of use metadata especially in the form of tagging. How might we use the other types of metadata to work through the noise of tags to discover the “real or useful” metadata or would that be disruptive in and of itself?
1. Metadata can have various definitions like it can be used to ‘tag’ an image, can be the file size of the digital file or something more complex like the subject matter of the resource or legal rights associated with its use. Considering these examples, is metadata the shredded form of the original data? If it contains the subject matter, then how can metadata be a descriptor to the original data?
ReplyDelete2. Various kinds of Schema are discussed such as the DC Schema, SEPIADES schema, the METS schema etc. Which of these is more efficient than the other and why? Can we use one of these or form universal schema that can be used in all databases?
3. The two types of metadata are implicit and explicit metadata. Implicit metadata is more useful to the administrators than the users. Is there a way to delete them when they are redundant or not used? Will this reduce the value of the data ? Can the system become more efficient by doing this as it frees up memory space?
1. Is metadata about metadata categorized/structured in the same way as traditional metadata?
ReplyDelete2. In consideration of intrinsic and extrinsic metadata -- is there ever an instance where metadata is both? If so, what does that look like, and how would it be classified?
3. I'm curious as to how metadata changes. For example, if information about an object is modified over time, is the original metadata forever associated with it? Are there formal processes or frameworks for addressing changing metadata information?
1. At one point, the author writes that “metadata invariably offers a selective or simplified description of a source”. Does this imply that metadata is often subjective? Is the inclusion of certain metadata dependent upon the repository?
ReplyDelete2. The author discusses how metadata is often held within databases but can, if decided, be embedded within the digital file itself. Can it be argued that one or other of these approaches is better or more effective? Doesn’t embedding within the file cause issues in that then the file changes every time you add new metadata, changing its original form, whereas storing it in a database allows the file to stay static while metadata can be added at any point and any time?
3. In discussing standards, the author differentiates standards base on administrative versus descriptive. They go on to say that “there have also been some attempts to create standards that are focused on particular purposes”. This seems like a very vague statement. Is it written as such because delving deeper is not within the scope of the paper, or if not, what exactly are these “particular purposes”? He cites PREMIS, which I’m more or less unfamiliar with.
1. I think that metadata is subjective. For archives and libraries,a repository decides the minimum number of fields that must be added to a catalog record. Additional metadata added could be added based purely on what the cataloger feels is important to add, and this can be impact by their own experiences.
DeleteI am also intrigued by metadata that is added by different users based on the different purposes of the object in its life cycle. For example, a person may take a photo of their mother and label it "Mom." Many years later, this photo could end up in a historical society, and it may be kept for historical reasons because the mom is standing next to a historical building. A cataloger there may re-title the photograph based on the name of the building, no longer really caring who the woman is or they might not even know. Here I think the metadata is subjective based on the reason the object is deemed important by different people. I am wondering if there is a way to embed in the object all of the metadata changes the object may have applied to it in its lifecycle.
1) Like many of our readings, this one called to mind some of the difficulties of language-based information description and retrieval when handling non-language-based resources. Some descriptions of their example, the Mona Lisa, are obvious (e.g. “woman” and “Renaissance”), but other potentially important or interesting information about the work may be less obvious, like the fact that it was commissioned by a wealthy man as a portrait of his wife, or the presence of a landscape in the background. How do metadata creators draw the line between vital information and the merely potentially useful? Does it vary by collection?
ReplyDelete2) The article points out some of the difficulties caused by “legacy” metadata schemas, in fields where pre-digital information retrieval systems have led to a schema that has since become entrenched even as its usefulness has waned. How is it possible to modernize metadata schemas while still accounting for the needs and expectations of the established community? Is it better to compromise and gradually evolve, to implement a new and unfamiliar schema that suits the information’s needs better, or to adhere completely to the possibly-outdated schema?
!. If cost was not an issue in the creation of descriptive metadata, would more necessarily be better? In looking at the Mona Lisa example, it seems that a very simplified approach is sufficient for a single information object and that too much descriptive metadata would make retrieval more difficult. What types of objects would benefit from a lot of descriptive metadata?
ReplyDelete2. The author compares HTML to XML saying that HTML is aimed at presentation while XML focuses more on meaning. The example given doesn't do much in the way of clarification. What does the author mean by this?
3. When discussing implicit metadata, the author states that at times some is extracted and stored elsewhere. The reasons he gives seem very valid but he says that it is typically not done. Why wouldn't this be done? Is it because of cost and time issues?
1. I have never heard of “resource discovery” metadata before this article. The article mentions that it is the metadata that helps us find a resource. I am wondering if this is a particular metadata field, or if this is more of a theoretical concept that can change with each search. For example, I could search for a book in a library catalog by its title. Another person may search by the author’s name. Does this make the title or the author the resource discovery metadata in each case, or are both in total the book’s resource discovery metadata?
ReplyDelete2. The author mentions that sometimes particular communities will “incorporate old-fashioned or ‘legacy’ approaches that may have worked well in a non digital environment, but are not as practical or useful in the digital world.” I am wondering what examples of legacy metadata exists in the archives and libraries fields. Are there aspects of archival metadata, such as a specific field or format of a finding aid, still included on digital finding aids, merely because they were documented in a certain way on paper, but might make sense to represent a different way now that the finding aid is digital? Were there metadata practices that were carried over into MARC from card catalogs, that doesn’t make the best sense given the digital aspect of MARC records?
1. " Metadata is usually structured in some way." Are there Metadata standards or uses that don't have any structure and just freely associate metadata with the object? It seems to me that doing this would not be very productive. Sure, the information is still stored and associated with the object but if it's not in some structured format I don't see how it could be useful.
ReplyDelete2. "As with the different levels, metadata schemas tackle this problem in different ways. Some will create sub-categories (e.g. Creator_OriginalWork; Creator_SurrogateImage); others, separate categories (e.g. Artist; Photographer; Scanning Technician); others, completely separate records for each layer." What is the advantage of keeping separate records for each layer? If someone is interested in a scanned image of the Mona Lisa I may not want t know who the scanning technician is but I would rather have the information and not need it in the case that I may want to know later down the road.
3. What are the advantages of storing metadata in a closed system like a database or within the actual file itself? I would think that storing metadata in an open and accepted format like XML so that if he institution that is curating these files has the most options available to them.
1. Under the section, “What is Metadata?” it says that metadata is often held within databases, but it can take other forms - it can just as easily be found embedded within the digital file itself." Does metadata take other forms other than being “data about data”? Also, does metadata only exist in print or online records? Could a person be metadata somehow? Ex.) Say a person is wearing a shirt to advertise something. Is that person essentially data showing where to find other data or resources?
ReplyDelete2. At what point does metadata end? Is it when it gets updated or becomes obsolete? Or is there metadata about when other metadata was created or destroyed and so on ad infinitum?
3. In talking about intrinsic and extrinsic metadata it says that extrinsic data is created by humans whereas intrinsic metadata is technical in nature. Since technology is created by humans doesn't that make intrinsic metadata extrinsic in nature as well?
1. The article briefly mentions that intrinsic metadata is recorded or collected with little or no human intervention, like a digital photo retaining information about the type of camera that took the picture. I remember this being an issue in the news recently, with photos taken with iPhone cameras and uploaded to Facebook/the Web making this metadata, including time and location that the picture was taken, available to tech-savvy users. The average non-iSchooler iPhone user probably wouldn't know that this information is being collected, but it can prove to be a huge security risk. How big a matter of professional ethics should it be for this information about metadata practices (meta-metadata, if you will) to be publicized about products?
ReplyDelete2. How possible might it one day be to make extrinsic metadata automatically available, without the high level of human involvement currently required? For instance, with OCR and other document reading technologies, would it be possible to write a program that reads documents in a collection, seeks particular collocations or patterns, and generates descriptive metadata about that document? (This may already exist; forgive my oblivion if that's the case.)
1. Metadata might be simply a free text description or set of keyword used to annotate or ‘tag’ an image. It is sure free text will mike metadata easy to generate, but will it cause the metadata hard to be understand or retrievable?
ReplyDelete2. Metadata aims to cover so many functions, like what it is, where the resource has come from, who owns it… But is it efficient? A metadata that might beyond the size of original data will increase the workload of database.
3. Metadata should serve communities and users, so it can take many different forms. The different standards will make metadata not compactible in other system. So should we create a universal standard for the metadata, or part of it?
1. Are pre-digital forms of descriptive information metadata? Are encyclopedias metadata? Dictionaries? Graffiti tagging? Deeds and titles? The Domesday Book?
ReplyDelete2. How has crowd-sourced/open metadata creation influenced professional library, archive, and museum practice? How has it affected academic theory on metadata? Will it be possible for information professionals to retain authority over metadata in an increasingly open-access, online information universe? How must we reconceptualize our roles and responsibilities as metadata authorities?
3. What, if any, programming language and/or software is being developed to "merge" various metadata standards into a larger, multi-coded metadata system (e.g. Dublin Core and METS)? What are the possibilities and pros and cons of such an impulse? Will it ever be possible - or desirable - to have one, all-encompassing, multi-multi-field meta-metadata schema?
1. It seems like metadata could be divided continuously as long as the data managers wish. For example, the location of "Mona Lisa" can be also divided into museum, city, state, and country.Some patrons might want to get to know these information, while others might find it useless. So,which aspects might be taken into consideration when we want to decide how detailed the metadata should be?
ReplyDelete2. The author claims that there are four categories or types of metadata: descriptive metadata, administrative metadata, structural metadata, and use metadata. Is it necessary and possible to collect all four types metadata for one document?
3. In the "Different communities and users" part, the author says that "We've just distinguished the end users of the collection from those managing the collection ( who are also "users")." What differences do these two kinds of users have when they are using metadata?
1. In this article the author states that there are different schemas of metadata that are created for specific purposes. These types of metadata include descriptive, administrative, structural, and use metadata. Since there are different schemas that focus on these different types of metadata then is it possible to have two different schemas of metadata that are applied to the same collection? If so then what are the benefits and drawbacks of doing this?
ReplyDelete2. In this article the author states that use metadata is metadata that is created by or about the users. We have seen that crowdsourcing is a viable method for getting users to create metadata to describe the contents of a collection. Would you consider this to be use metadata or descriptive metadata since it is useful to describe the collection?
3. In this article the author states that there are three different places that metadata about an object can be found. It can be found inside the information object, in a separate XML file, or in a database. Which of these three places do you think is the best place to store metadata and why? Which of these places do you think is a bad place to store metadata and why?
1) If an advantage of metadata is to share the metadata with other similar collections, how does choosing to describe items according to different layers affect the share-ability of the metadata? If one institution wants to create sub-categories, another separate categories, or another separate records for each layer are these records useful to other institutions or users trying to explore the collection?
ReplyDelete2) In the section “Where does metadata come from?” it is stated that there is intrinsic and extrinsic metadata, the difference being whether a human or machine created it. Isn't automatic metadata ultimately also created and associated with a resource by the humans who created the programs that automate the metadata? This was simple and obvious metadata based on file characteristic, formats, and origins of the file, but is there an essential role for information professionals to be involved in the design of programs that produce automatic metadata and to what level should they be involved?
3) Metadata can be stored in different places depending on what it is and what it is used for, some stored with the file other just through association. What are the advantages to combining or separating administrative, descriptive, structural, preservation, and use metadata and does it depend on the resource you are describing, a random photograph vs. the Mona Lisa?
Q1. The article mentioned the image metadata and set the portrait Mona Lisa as an example to illustrate that metadata is usually structured in some way. I am wondering who is responsible for selecting the metadata elements that associated with an image? When factors do they need to consider when making the decisions?
ReplyDeleteQ2. In the Different purposes section, the author states that metadata can serve different purposes, like help us find resource or tell us what it is. So I want to know which purpose is the one that promote the emergence of metadata?
Q3. The author talked about the metadata applied in different communities and facing different users. Is there a role who is exclusively in charge of metadata management in any kinds of communities? How do they tackle the outdated metadata?
1. I can imagine where one comes across “resource discovery” metadata, or “descriptive" metadata but where does one come across “provenance and rights” or “technical” metadata?
ReplyDelete2. It seems like the more one adapts a metadata schema to a particular user community, the less potential it has to become more universally searchable as it won’t be compatible with any other metadata schemata. Isn’t this problematic?
3. Can any metadata schema be written in XML?