Dcer Logistics And Formats ( 2f ) Example Metadata File

 

#format text

===================================================================

D C E R Metadata Template

===================================================================

This file is a plain Java properties file that holds the metadata

for this experimental dataset. All of the fields are commented so

that it can serve as a template.

——————————————————————-

Required Properties

——————————————————————-

The properties in this section should always be provided for all

assignments, all the time.

——————————————————————-

A title for this dataset, provided by the contributor. Like a paper

title, it should be as descriptive as possible. However, titles are

not presumed to be unique.

—–

dataset.title = Objectdraw Opinion Survey

A comma-separated list of one or more languages in which this dataset’s

files are written. Values should conform to RFC 4646 (see

http://www.rfc-editor.org/rfc/rfc4646.txt). If no value is

given, then “en” is assumed.

#

Individual files within the dataset should use one language only, but

multiple translations within the dataset are fine. Note that the

first entry in this list is considered the “primary” language for

this dataset, and it is expected that all supporting files will be

available in that language. Other languages in the list would

then imply that translations for some or all of those files in the

given language would also be provided as part of this dataset.

—–

dataset.languages = en

An optional comma-separated list of keywords or tags provided by

the contributor.

—–

dataset.keywords = objectdraw, JUnit, TDD, test-first coding, \

test-driven development

Basic contact information. Can be repeated as many times as necessary

with different numeric suffix to cover the number of authors/contributors

involved. Only one author is required, and only the e-mail is required.

The other author fields are actually optional.

—–

author1.name.first = Stephen H. author1.name.last = Edwards author1.email = edwards@cs.vt.edu author1.url = http://people.cs.vt.edu/~edwards/ author1.phone = +1 540 231 5723 author1.institution = Virginia Tech author1.department = Dept. of Computer Science

An optional element that gives one of the following values, characterizing

the general nature of the data available in this dataset:

#

+ transcript : The dataset consists of interviews, transcribed video,

“think aloud” protocols, or other scenarios where the

primary data is transcribed text.

+ survey : The dataset consists of answers to some kind of survey.

+ exercise : The dataset consists of student performances on an

assigned exercise, test, quiz, etc., other than a

programming exercise.

+ programming : The dataset consists of student performances on a

programming exercise or activity.

#

For datasets that cover multiple categories, a comma-separated list

can be used.

—–

dataset.type = survey

The copyright holder and date. This can be automatically filled in

when the dataset is submitted (default == current year and

contributor).

—–

copyright = (c) 2008 Virginia Tech Department of Computer Science

The name of the license under which rights to use this dataset

are granted. If omitted, the assumption is “all rights reserved

by the copyright holder”. When a new dataset is uploaded, we

can suggest this creative commons license as the default. We can

also provide a drop-down list of the most common choices, including

the creative commons choices and approved OSI open source licenses.

—–

license = Creative Commons Attribution-Noncommercial 3.0 United States License

The URL under which the full text of the license is available on-line.

—–

license.url = http://creativecommons.org/licenses/by-nc/3.0/us/

——————————————————————-

Recommended Properties

——————————————————————-

The properties in this section are expected for all assignments,

but stingy or lazy contributors may still fail to provide them

(we’ll bug them, but we won’t force them).

——————————————————————-

A brief, one- or two-sentence description of what this dataset contains.

This will be the short description other researchers will first see when

searching. Keep your entry to 30 words or less.

—–

synopsis = An opinion survey given to CS1 students after completing a \

course that required unit testing of GUI assignments written using \ objectdraw.

——————————————————————-

Subject Population

——————————————————————-

Number of subjects involved in experiment.

—–

subjects.count = 41

Optionally, identifies the academic course(s) involved in the

experiment. Can be a comma-separated list. Where possible, complete

course numbers and titles should be provided, although contriutors can

also use generic terms (like “CS1”) or tags (like “introductory”) or

any other appropriate label they feel will be recognizable by others.

—–

subjects.course = CS1, CS1705: Introduction to Object-oriented Development I

An optional comma-separated list of keywords or tags intended to identify

the knowledge content that is the focus or context of the experiment,

where appropriate.

#

If possible, CC2001 (or newer) knowledge units can be used in this

spot, although contributors are also free to use more generic terms,

including any other appropriate labels they feel will be recognizable

by others.

—–

subjects.content =

——————————————————————-

Optional Properties

——————————————————————-

Optional properties are not required. We’ll give contributors a

chance to provide them, but won’t nag. We expect that many

assignments won’t provide some or all of these.

——————————————————————-

——————————————————————-

Citations

——————————————————————-

We want to be able to include citations for papers published using

or referring to this dataset. There are many different forms, however.

My suggestion is to (a) prompt for a real URL, and then (b) take

any human-readable form and represent it as a plain string. After

all, it is relatively unlikely that anyone is going to be automatically

building bibliographies directly off such a repository’s optional

citation info, and as long as it is human-readable, users should be

able to convert it to an appropriate format. Especially if a real

url is provided. So anything from a plain text formatted reference to

bibtex or anything else seems reasonable.

citation.1 = … citation.1.url = http://… citation.2 = … citation.2.url = http://…

——————————————————————-

Repository Properties

——————————————————————-

Repository properties are provided automatically

by a third party, typically when an assignment is uploaded to some

service. These properties are not intended for direct editing by

end users or contributors, but are provided in downloaded assignments

for informational purposes, and to aid in tracking the relationships

between modified/adapted versions of assignments and their original

ancestors.

——————————————————————-

A globally unique identifier for this dataset, generated by the

repository or service to which the assignment is uploaded. The property

also includes a rudimentary checksum-style hash code for error detection

and tamper resistance.

—–

repository.id = dcer.org:12345 de23

A generated version number for this dataset. We can just use

integers. Any of the registered authors can upload a mod, which

is recorded as another version.

—–

repository.version = 2

A publicly accessible URL where people can view information about this

dataset, typically provided by some repository/service to which this

dataset has been uploaded. Can this be the same as the repository id?

Maybe it should be

—–

repository.url = http://dcer.org/dataset/12345

A human-readable change history summary for this dataset. If the

repository/service hosting this dataset allows contributors to

briefly describe changes when they upload modified/updated datasets,

then a consolidated presentation of the change history associated with

this particular version can be generated here for informational purposes.

—–

repository.history =

——————————————————————-

Additional Properties

——————————————————————-

I’ve only put in the basics so far. If we go through Jan Erik’s

merging of our metadata suggestions, we can add more as we see

fit.

#

Contributors can add any properties they want as well.