24
14

There's an increasing variety of data available as Linked Data coming from a range of different sources. I'm wondering what indicators we might use to judge the "quality" of a dataset.

Clearly quality is a subjective thing, but I'd be interested to know what factors people might use to indicate whether a dataset was trustworthy, well modelled, sustainable, etc.

I've marked this as a community wiki question so we can collaborate on a list.

Summary of some of the key points from these answers and discussion on the LOD list:

  1. Accuracy - are facts actually correct?
  2. Intelligibility - are there human readable labels on things?
  3. Referential correspondence - are resources identified consistently without duplication?
  4. Completeness - do you have all the data you expect?
  5. Boundedness - do you have just the data you expect or is it polluted with irrelevant data?
  6. Typing - are nodes properly typed as resources or just string literals?
  7. Modeling correctness - is the logical structure of the data correct?
  8. Modeling granularity - does the modeling capture enough information to be useful?
  9. Connectedness - do combined datasets join at the right points?
  10. Isomorphism - are combined datasets modeled in a compatible way?
  11. Currency - is it up to date?
  12. Directionality - is it consistent in the direction of relations?
  13. Attribution - can you tell where portions of the data came from?
  14. History - can you tell who edited the data and when?
  15. Internal consistency - does the data contradict itself?
  16. Licensed - is the license for use clear?
  17. Sustainable - is there a credible basis for believing the data will be maintained?
  18. Authoritative- is the provider of the data a credible authority on the subject?

Further sources:

This question is marked "community wiki".

asked 24 Jun '10, 23:52

ldodds's gravatar image

ldodds
1.8k1314
accept rate: 32%

edited 16 Apr '11, 15:04

zazi's gravatar image

zazi
3.4k1213

1

Can you clarify whether you're seeking to find a good set of factors which publishers can use in order to refine and improve their datasets - or - looking to find a way to rank, rate and judge any old data set on the web as if to say one is somehow 'better' than another?

(25 Jun '10, 08:29) Nathan Nathan's gravatar image

I was thinking more about how to rate and judge any dataset on the web, although I suspect that the two issues are closely related. If there are some good indicators of quality then publishers may want to (re-)think how they expose data.

(25 Jun '10, 13:52) ldodds ldodds's gravatar image

I do wish that a discussion about Linked Data would have any links (http or otherwise) therein be live... The original question's link to the LOD post should not be a literal string above (nor does it appear that it was intended to be a literal, by the visible markup, so perhaps that's correctable?). And it won't be, here -- http://lists.w3.org/Archives/Public/public-lod/2011Apr/0140.html .

(15 Apr '11, 09:42) TallTed TallTed's gravatar image

Ted, it's a wiki, so perhaps you could figure out how to fix it?

(16 Apr '11, 12:19) Ian Davis Ian%20Davis's gravatar image

Ian - I did look around for a link to make such edits ... but though the question is marked as "community wiki" I see no such link. Maybe my karma is too low to be granted such privilege, or maybe I'm just blind. Perhaps you could provide an actual link and/or guide me to where it should be on this page I'm looking at, instead of just being cleverer-than-thou?

(18 Apr '11, 11:06) TallTed TallTed's gravatar image

12next »

For starters, I think we can all agree that at its core the measure of data quality is subjective and that "beauty is the eye of the beholder": the quality of data is the measure of a dataset's fitness for use in the specific application. Indeed, in his 2005 white paper Principles of Data Quality [1] Arthur D. Chapman writes,

Data quality is multidimensional, and involves data management, modelling and analysis, quality control and assurance, storage and presentation. As independently stated by Chrisman [2] and Strong et al. [3], data quality is related to use and cannot be assessed independently of the user. In a database, the data have no actual quality or value [4]; they only have potential value that is realized only when someone uses the data to do something useful. Information quality relates to its ability to satisfy its customers and to meet customers’ needs [5]

Chapman enumerates the factors contributing to fitness-for-use, citing Redman [6]:

  • Accessibility
  • Accuracy
  • Timeliness
  • Completeness
  • Consistency with other sources
  • Relevance
  • Comprehensiveness
  • Providing a proper level of detail
  • Easy to "read"
  • Easy to "interpret"

Each of these factors is fundamentally subjective, even if mechanisms exist within particular domains to take their measure "objectively." Indeed, in some domains such ratings might only be done by humans, either through voting mechanisms or by individual reviewers.

The greater linked data community needs to develop vocabulary terms for expressing metrics for data quality -- consider the ten points above -- and then within individual communities develop agreed-upon means to determine those values. Arguably this is the "Dublin Core" approach to the problem, in the sense that terms like completeness or consistency would be reused across domains with inherently different domain-specific meanings, but such reuse would facilitate consumers from other communities choosing datasets. "The physics community says this dataset is accurate, by their measure.

Some of these factors might be calculated dynamically, based on the consumer's context. An example of this is relevance, which could be interpreted as equivalent to a recommendation.

References (from Chapman):

  1. Chapman, A. D. 2005. Principles of Data Quality, version 1.0. Report for the Global Biodiversity Information Facility, Copenhagen.
  2. Chrisman, N.R., 1991. The Error Component in Spatial Data. pp. 165-174 in: Maguire D.J., Goodchild M.F. and Rhind D.W. (eds) Geographical Information Systems Vol. 1, Principals: Longman Scientific and Technical.
  3. Strong, D.M., Lee, Y.W.and Wang, R.W. 1997. Data quality in context. Communications of ACM 40(5): 103-110.
  4. Dalcin, E.C. 2004. Data Quality Concepts and Techniques Applied to Taxonomic Databases. Thesis for the degree of Doctor of Philosophy, School of Biological Sciences, Faculty of Medicine, Health and Life Sciences, University of Southampton. November 2004. 266 pp.
  5. English, L.P. 1999. Improving Data Warehouse and Business Information Quality: Methods for Reducing Costs and Increasing Profits. New York: John Wiley & Sons, Inc. 518pp.
  6. Redman, T.C. 2001. Data Quality: The Field Guide. Boston, MA: Digital Press.
permanent link
This answer is marked "community wiki".

answered 12 Jul '10, 13:47

olyerickson's gravatar image

olyerickson
413
accept rate: 0%

"data quality is related to use and cannot be assessed independently of the user" + "Information quality relates to its ability to satisfy its customers and to meet customers’ needs": That's why, I try to establish information service quality ratings from information service quality rating agencies, where the user/customer can select from and use ratings from agencies, which he/she trusts. Thanks a lot for this insightful comment!

(12 Jul '10, 21:35) zazi zazi's gravatar image

We could define LODrank as a pageRank-like measure that was a function of the number of links to/from other LOD datasets weighted by their LODrank. Alternatively, it might divided by the number of linkable instances in the collection, so that large datasets did not have an advantage. This metric addresses quality indirectly, depending on the fact that other LOD collections will be linked to a dataset if it is useful, which depends on how much data it exposes, how it's encoded, the data's quality, etc.

permanent link
This answer is marked "community wiki".

answered 25 Jun '10, 06:20

Tim%20Finin's gravatar image

Tim Finin
701128
accept rate: 25%

edited 25 Jun '10, 13:08

I like the idea of using linking as a proxy for quality, just as we do for HTML pages. I suspect there's the same potential for "gaming" of that system as there is for the open web, i.e. creating reciprocal links between datasets of dubious quality/provenance.

(25 Jun '10, 15:49) ldodds ldodds's gravatar image

+1 for LODRank ; could possibly factor in blank nodes as some sort of negative/neutral impact.

(26 Jun '10, 16:20) Ed Summers Ed%20Summers's gravatar image

It's an interesting metric, though, even without link farmers gaming the system, I think you'd have to be careful about assumptions you can make about why the links are created. eg: I mostly create owl:sameAs links to dbpedia because they provide good identifiers; in many cases I won't have made any judgment at all on the accuracy or modelling of the dbpedia data.

(12 Jul '10, 23:22) kwijibo kwijibo's gravatar image

@ Idodds ... but like with PageRank the reciprocal link networks won't have much LODRank to share amongst themselves ... so it won't kill the system, it would more likely be a minor nuisance.

(25 Dec '10, 09:47) Ankur Ankur's gravatar image

Hi Tim, just to make you aware, a ranking algorithm as the one you described is explained and evaluated in [1,2]. This is an extension of the DING algorithm developed in collaboration with Michael.

[1] http://renaud.delbru.fr/doc/pub/eswc2010-ding.pdf [2] http://km.aifb.uni-karlsruhe.de/ws/semsearch10/Files/deri.pdf

(09 Feb '11, 23:16) Renaud Delbru Renaud%20Delbru's gravatar image

Brilliant question!

Indeed, there is a lot to do in this area. For a start, yes, I agree with Tim re the ranking (see also our LDOW09 paper "DING! Dataset Ranking using Formal Descriptions", essentially utilising voiD to do the job, now implemented in Sindice).

However, I understand that there are several aspects (as suggested by Leigh) one has to take into account:

  1. Quality of the raw data (the crap-in-crap-out syndrome); tools such as Gridworks can help here
  2. Quality of the RDFisation process (IN:some raw data, OUT: RDF using some vocabularies); here the main issue, I think, is the lack of (good) vocabs and also the discovery process (lot of people reinventing the wheel rather than reusing mature vocabs, where available)
  3. Quality of the interlinking process (IN: RDF dataset, OUT: RDF dataset + links to other datasets); here the main issue IMO is the lack of interlinking frameworks. The only reliable, usable one I'm aware of is Silk, though limited to owl:sameAs links.
  4. Quality of making more data explicit (aka turning data into information or simply put: reasoning) - there are ongoing efforts such as Scalable Authoritative OWL Reasoning for the Web, but that's not domain of expertise, so I better stop here ;)

An accompanying activity is the Pedantic Web group, providing input and support re the quality of Web data.

Thanks, Leigh, for setting the ball rolling - we should definitely continue working on this issue on a collaborative basis.

As an aside: in September 2010 we'll kick-off a project, which mainly focuses to address this issue. If you're interested in details, ping me.

permanent link
This answer is marked "community wiki".

answered 25 Jun '10, 08:01

Michael%20Hausenblas's gravatar image

Michael Haus...
78426
accept rate: 15%

Glenn McDonald provided an excellent list of 15 ways to think about data quality:

  1. Accuracy - are facts actually correct?
  2. Intelligibility - are there human readable labels on things?
  3. Referential correspondence - are resources identified consistently without duplication?
  4. Completeness - do you have all the data you expect?
  5. Boundedness - do you have just the data you expect or is it polluted with irrelevant data?
  6. Typing - are nodes properly typed as resources or just string literals?
  7. Modeling correctness - is the logical structure of the data correct?
  8. Modeling granularity - does the modeling capture enough information to be useful?
  9. Connectedness - do combined datasets join at the right points?
  10. Isomorphism - are combined datasets modeled in a compatible way?
  11. Currency - is it up to date?
  12. Directionality - is it consistent in the direction of relations?
  13. Attribution - can you tell where portions of the data came from?
  14. History - can you tell who edited the data and when?
  15. Internal consistency - does the data contradict itself?
permanent link

answered 13 Apr '11, 05:07

Ian%20Davis's gravatar image

Ian Davis
2.3k1414
accept rate: 13%

edited 13 Apr '11, 05:11

1

+1, but should maybe add Dave's additions... Consistency of modelling, Licencing, Sustainable, Authoritative... http://lists.w3.org/Archives/Public/public-lod/2011Apr/0145.html

(13 Apr '11, 07:06) Signified ♦ Signified's gravatar image
2

yep, i didn't want to conflate the two posts but this question is a community wiki so we could just edit it, in fact, that's what I'm going to do!

(13 Apr '11, 07:34) Ian Davis Ian%20Davis's gravatar image
1

Awesome! Now to quantify them... :/

(13 Apr '11, 18:47) Signified ♦ Signified's gravatar image
1

And all of which will eventually be summarized into one question: Is the dataset useful? :)

(28 Feb '13, 13:21) fadirra fadirra's gravatar image

In my opinion, data quality is not necessarily subjective. Imagine valid combinations of cities and countries or accurate population values. Who defines these quality measures? Surely not an individual. These examples are rules that have been derived from public knowledge or stated by natural circumstances. So besides individual requirements from data consumers, the following things may at least also be sources for data quality rules:

  • real-world phenomena (e.g. city/country combinations)
  • organizational policies (e.g. all TV's in my data must have a screen size)
  • legal regulations (e.g. all groceries must have an expiration date)
  • IT-needs (e.g. URI's must be dereferencable)
  • Standards (e.g. the syntax of xsd:dateTime or ZIP codes)
  • task requirements: e.g. population data for all populated places must be complete to calculate the world population.

As part of my PhD, I am currently investigating the quality of Semantic Web data sets on instance-level. I have also published a data quality constraints library at http://semwebquality.org/ontologies/dq-constraints# which may be used in conjunction with SPIN ( http://spinrdf.org/ ) to identify potential data quality problems. The constraints do not directly restrict the openess of the web, since it is up to the data owner/provider whether the instances with potential data quality problems shall be cleansed. The constraints shall rather help to identify incorrect or suspicious data and raise transparency about the quality state of the Semantic Web data sets in first place. If you are interested in this kind of quality assessment, please see my publications and/or contact me.

permanent link
This answer is marked "community wiki".

answered 25 Oct '10, 18:52

Christian%20F%C3%BCrber's gravatar image

Christian Fü...
11
accept rate: 0%

As Michael said, "Brilliant question!" ;)

In addition to the mechanisms Michael suggests -- each of which has the advantage of being automation-friendly -- I think we must also ask how to introduce social mechanisms (crowdsourcing, etc) and provenance infrastructure into the mix.

As we learned most recently from IPAW2010, provenance can be an important indicator of "quality." The methods used of course depend on the consumer and the application; an otherwise functionally valid and "relevant" dataset might be rejected because there is insufficient and/or irrelevant provenance metadata attached.

Reviews and rankings gathered through social mechanisms and communicated via linked data principles can be indicators of quality to the community. I'm a firm believer that every presentation of a dataset, whether via a SPARQL endpoint or (simply) a CSV download target, should be presented in a social context that accommodates commentary. "Very clean!" "The units are inconsistent!" etc. There are many exemplars out there, ranging from myExperiment to iGoogle Gadgets to the Google Apps Marketplace.

Ultimately, provenance and the "social machine" will be inextricably linked. Issues of scale will require many services to depend on crowd-sourced data scrubbing and transformation; these services will in turn rely on provenance standards that enable third-party applications and services -- upstream in the value chain --- to indicate the sourcing and processing of the datasets they provide and the parties responsible for that processing.

Data "quality" is in the eye of the beholder, and provenance data is part of what that eye needs to see...

permanent link
This answer is marked "community wiki".

answered 25 Jun '10, 12:32

olyerickson's gravatar image

olyerickson
413
accept rate: 0%

1

In that case, does there also need to be a mechanism for assessing the quality of data providers - do they / don't they take the comments from the crowd and fix the data in a timely way? Does it stay fixed in the next release after? etc.

(25 Jun '10, 13:10) Ian Ian's gravatar image

I think that's exactly the case. How to indicate it and how to generate it are two separate questions, however. Indicating it might simply be a case of vocabulary, but the mechanisms for generating it might vary from voting schemes (like this portal, or SlashDot) to arbitrary assertions.

Related to this is the ability to query the provenance graph of a dataset. This was featured in a few IPAW2010 papers and is very intriguing; imagine being able to query whether certain services were used to process a dataset, etc.

(25 Jun '10, 13:20) olyerickson olyerickson's gravatar image

At a high level, the main drivers for me are: how easy it is to find the information I'm looking for; and conversely, how hard is it to find wrong information. From this point of view, and concentrating on the semantic web portion of the picture, we want to figure out measures and metrics for how well a particular (RDF/linked data) dataset or set of datasets satisfies these drivers.

I can't see how a naive count of the number of links between datasets can offer much of a measure of how easy it is to find what you're looking for, nor how difficult it is to find wrong information. There's probably some deep scale-free characteristic (c.f. the six degrees of separation meme) suggestive of a perfect balance between having no links -- which feels like it should be bad -- to having a complete graph with links everywhere -- which feels like it would be a needle+haystack and equally bad. There'd have to be some evidence to back up that sort of conjecture though, and wouldn't there always be ways to "cheat" and just inject enough random links to optimize this metric?

Given that most of the time we're going to be working with existing data and models, and transforming that data to linked data and ontologies, we need to consider metrics which can factor out the quality metrics of the underlying data and instead show the added benefit of the semantic web approach.

One way to look at things is to consider the methodology used in transforming data to linked data and making it available, and figure out ways to measure how well this process helps with the overall aim of making it easier to find the right information and harder to find wrong information.

The methodology I tend to use starts off by re-representing the underlying data in RDF and OWL fairly simplistically and as "one to one" as I can manage. By "one to one", I like to think that the transformation so far would only have made things more explicit, and not removed any information, so it should be possible to reverse the transformation and get the original data back. Perhaps one measure at this stage would be to show how easy it is to get back the original data e.g. using the linked data API?

A next (notionally at least) step is adding value by showing how the data can be interpreted in existing ontologies/schemas. Actually, this tends to happen at the same time as transforming the data anyway, but since we're arguably adding value at this step, we should consider it separately. For me, thinking of this as a separate step highlights another of the metrics we tend to focus on: re-use of existing vocabularies.

As a community, I think we tend to focus on this re-use of vocabularies and frown upon creating new vocabularies, to the point where we end up trying to fit square pegs in round holes again. In my experience, most existing models have quite specific nuances and requirements and there should be nothing wrong in making the underlying models explicit and self consistent using new vocabularies or ontologies. At the same time, where there is obvious alignment, it can only help understanding to link to and re-use existing vocabularies, but that re-use can often just as well be via specialization, e.g. my:Person rdfs:subClassOf foaf:Person.

Rather than consider it bad to coin new vocabulary terms, we could instead focus on how "tight" these vocabularies are at constraining the data to be self consistent, and when linked to other vocabularies, constraining the data to be consistent and meaningful with those linked vocabularies. As the Wang and Strong paper above points out, most datasets contain non-trivial amounts of wrong data, so perhaps another good metric would be how much of this bad data we could surface along the way. And if there is no bad data, we could figure out some kind of fuzz-testing regime.

Anyway, that's just one way of creating SW datasets, but it helps me keep a track of some of the characteristics I think are important.

I also strongly agree with olyerickson that including provenance along with your data is incredibly important, which also raises the issues of trust and proof as we eventually move up the layer cake :)

permanent link
This answer is marked "community wiki".

answered 25 Jun '10, 16:19

ajtucker's gravatar image

ajtucker
1083
accept rate: 50%

+! from me, especially the naive count reference! analyse each thing at a granular level taking everything in to account and then if you must try to get some bearing on the full 'dataset' - why you'd want to rank datasets is beyond me though

(25 Jun '10, 17:05) Nathan Nathan's gravatar image

I don't see it as a question of ranking datasets, rather of helping dataset publishers do it as 'well' as possible. So to help people publish data in a useful form, we need some kind of clear criteria on how well they have done it.

(12 Jul '10, 12:15) billroberts billroberts's gravatar image

Re. defining data quality (information qualitity) statements, the Info Service Ontology (1,2,3,4) could maybe help, because it offers a hook for defining information service quality ontologies, where information service quality ratings could be based on. A dataset could also be seen as information service (see the definition of the term 'information service'(5)). Hence, it might be of interest for dealing with quality ratings, which might also come from different information service quality rating agencies.

Cheers,

Bob

(1) The Info Service Ontology specification
(2) The concepts and relations of the Info Service Ontology
(3) A proof-of-concept example of the Info Service Ontology
(4) The blog of the Info Service Ontology
(5) Definition of the term 'information service'

permanent link
This answer is marked "community wiki".

answered 25 Jun '10, 12:05

zazi's gravatar image

zazi
3.4k1213
accept rate: 13%

1

A good classification of data quality/information quality is described in Beyond Accuracy - What Data Quality Means to Consumers (Wang & Strong, 1996).

(25 Jun '10, 12:17) zazi zazi's gravatar image

So in terms of the Info Service Ontology I guess I'm asking what those additional properties/descriptions should be. There's also the question of who produces them, how, etc.

(25 Jun '10, 15:52) ldodds ldodds's gravatar image
  1. I think there should be a general objective description of an information service, which categorizes this information service re. its purpose/content. Upon other terms, somehow similar as it is already possible for datasets using void:Dataset and dcterms:subject for associating (SKOS) categories, e.g. from DBPedia + defining a main subject, information service type and information service contributor type (please tell me further properties for describing an information service). Such a description should help to choose the (hopyfully) right or a good information service, which delivers data
(25 Jun '10, 21:51) zazi zazi's gravatar image
  1. With the concept is:InfoServiceQuality is there a possibility to define different information service quality ontologies. On the basis of such an ontology specification, information service quality rating agencies are able to rate a specific information service. The specific information service quality ontologies could differ in their complexity and purpose/target group. The goal is to inform a customer (a human being or a machine) of an information service about the quality of this information service. It enables the opportunity to select a information service by its quality rating.
(25 Jun '10, 21:58) zazi zazi's gravatar image
  1. An information service quality agency could be at least every foaf:Agent. However, I think that quality ratings, which come from official and/or well established organizations (this may differ re. the main subject/domain of an information service), would get a higher trust value from its 'consumer community' as a rating from an unknown private person. Although, the majority tends more and more into the long tail. That means, ratings from 'unknown private persons', which are maybe somehow popular in their milieu, will also get a higher trust value in their domain and the information service
(25 Jun '10, 22:05) zazi zazi's gravatar image

... would maybe select upon their information service quality rating. Finally, trust and information quality is always somehow subjective. That means we need ontology specifications, which are able to model all these dependencies and relations for personalisation. One task could maybe, to summarize all information service quality ratings to one single, somehow objective information service quality rating. I hope that I could answer you questions a bit and clarify the scenario. My comparing scenario is always the stock exchange market with stock ratings from finance rating agencies.

(25 Jun '10, 22:13) zazi zazi's gravatar image
showing 5 of 6 show 1 more comments

Here are 6 subjective factors that affect the quality of Linked Data:

  1. Accuracy
  2. Navigability
  3. Access Protocols (they should be platform independent e.g. HTTP)
  4. Data Representation Formats (the more the merrier)
  5. Provenance
  6. Change Sensitivity (Freshness).

Basically, this is why I say you can look at a unit of Linked Data value in the same way you look at a Cube of Sugar :-)

Note, this is about the data not the place serving up the data, or the data consumption context halo.

Kingsley

permanent link
This answer is marked "community wiki".

answered 02 Jul '10, 16:17

kidehen's gravatar image

kidehen
1531
accept rate: 6%

1

I think two of your items are explicitly about the serving of the data -- access protocols and formats. They don't have anything to do with the data quality, just the means of publishing/dissemination

(05 Jul '10, 13:31) ldodds ldodds's gravatar image

Re. "Place" I mean: not about the platform (e.g. Virtuoso or Virtuoso+ODS). Instead its about Data within or inside the Web medium which inherently delivers:

  1. Platform independence (i.e. not operating system specific)
  2. Data Representation Formats are negotiable.

My response is about HTTP and QoS factors affecting Data within HTTP Networks.

The Context of consumption and the HTTP compliant platforms are out of scope.

(12 Jul '10, 14:24) kidehen kidehen's gravatar image

It seems to me that @kidehen's 6 points (perhaps with a bit of label massage) are categories into which @ldodds's suggested 18 might be sorted.

For instance, @kidehen's #5, Provenance, seems to me to include @ldodds's #13, Attribution, #14, History, #16, Licensing, and #18, Authority. (I've adjusted these labels to be more consistent with each other.)

(15 Apr '11, 09:49) TallTed TallTed's gravatar image

I think this question would be easier to answer if we knew more about how the "quality" is to be used — as advice to a human or as filtering to mechanistic processing?

Leigh said it all when saying the quality of a source is subjective, I trust the BBC more than Fox News, however this question maybe better answered in terms of the accuracy of individual statements made, and indeed the subjectivity of the statement:

  1. the film is 91 minutes long
  2. this is a legal copy of the film
  3. the film is rated 5 stars

Depending on who asserted any of the above I might mechanistically accept 1, take 2 from a trusted source, but always take 3 with a pinch of salt.

And to cap it all, when it comes to humans there are too many cognitive biases which then come into play. A machine may tell me that on paper, a Dell is a better buy for me, but I may still really want that shiny-shiny Apple, and will use the facts to post rationalise my emotional needs. Your metrics say Fox is more accurate, but I will always prefer to put my trust in the BBC regardless because I don't like them.

What I suspect is behind this question is if 20,000 people claim a film is 102 minutes long, and one person says no, it's 91 minutes, it would be nice if you could use "quality" to work out who to trust. But then just because you are trustworthy doesn't prevent you from being wrong. I imagine a lot of trustworthy people blindly retweeted "Back To The Future Day", and it appeared on many trustworthy web sites.

In the LOD I'd imagine the answer is to make statements against statements regarding their provenance .. citations if you will, but then those citations themselves are all subject to the rambling worries listed above.

I guess what I'm saying this is the Semantic Web Kobayashi Maru test.

permanent link
This answer is marked "community wiki".

answered 12 Jul '10, 12:54

psd's gravatar image

psd
1798
accept rate: 16%

That's why, I propose information service quality ratings from information service quality rating agencies. This enable the issue, you've described. You would then choose BBC as information service for your application/ knowledge base, rather than Fox or some thing else, because you said that you trust the BBC.

(12 Jul '10, 13:31) zazi zazi's gravatar image

I actually said " I trust the BBC more than Fox News" which is subtly quite different to "trust the BBC". It may just be the wording, but "information service quality rating agencies" doesn't sound very appealing.

(13 Jul '10, 15:40) psd psd's gravatar image
Your answer
toggle preview

Follow this question

By Email:

Once you sign in you will be able to subscribe for any updates here

By RSS:

Answers

Answers and Comments

Markdown Basics

  • *italic* or _italic_
  • **bold** or __bold__
  • link:[text](http://url.com/ "title")
  • image?![alt text](/path/img.jpg "title")
  • numbered list: 1. Foo 2. Bar
  • to add a line break simply add two spaces to where you would like the new line to be.
  • basic HTML tags are also supported

Question tags:

×262
×64
×5

question asked: 24 Jun '10, 23:52

question was seen: 14,097 times

last updated: 04 Mar '13, 03:10