New tag proposals from iterating.org (1/2)

[Justin B Rye]

Enrico Zini wrote:
> the people at iterating.org have been working at a mapping between their
> categories and the debtags ones.  Attached is an opendocument
> spreadsheet with the draft mapping they have produced.

Well, many thanks to them (and you of course).

> In this post I go through all the items, extracting new tags for the
> debtags vocabulary and commenting the mapping here and there.  I invite
> other readers to quote me and provide their comments: there's a lot of
> good insight and proposals in the lines below.
> 
> "Teaching Tools"
> 	can be mapped to use::learning
> "E-Books"
> 	can be mapped with role::data
> 	[we had something for books like anarchism and cvsbook: did it
> 	 just end up as role::data?]

(No true anarchist would be using a centrally-imposed official tag
classification system anyway!)

> "Home Automation"
> 	no mapping.  We need it for packages like bottlerocket and wmx10

What have we been doing with "Section: electronics"?
 
> "Board Games"
> 	why game::adventure and game::demos as extra tags?
> "Card Games"
> 	same: why game::platform as extra tag?
> 	I guess I need more info on what is the "extra tag" column.
> "Console-based Games"
> 	in the website this lists software for game consoles.  I don't
> 	think we have any in Debian

(Oh, *that* sort of console-based.)

> "First Person Shooters"
> 	tag missing: PROPOSED: game::fps
> "Real Time Strategy"
> "Turn Based Strategy"
> 	we have game::strategy; iterating.org proposes
> 	game::real-time-strategy.  Maybe we can have
> 	game::strategy:real-time and game::strategy:turn-based ?

game::strategy:* seems a better idea than game::time:*, and dividing
RTS from turn-based makes some sense, but I'd vote for avoiding
sub-hierarchies in sets this small.  We should also be cautious
about basing our terminology on the latest popular categories of
game when our archives are mostly stocked with stuff like ninvaders.
 
> "Event Automation Tools"
> 	no mapping.  We need it for packages like cron, anacron,
> 	fcron...  PROPOSED: admin::scheduling

Only the clock-based ones?

> "Job Scheduling Tools"
> 	it looks like the same as "Event Automation Tools": in the
> 	website there are no packages, so I don't know what it refers
> 	to.

Non-clock-based things?  Except that all that sort of thing tends to
be done via run-parts directories in each daemon etc, not by some
central scheduler package.  Or maybe they mean "gato"?

(Roll on the day when we replace all of this stuff with something
like upstart.)
 
> "Output Management Tools"
> 	I don't understand what this means

Sounds like a euphemism for paper-shredders.  No idea.

> "Performance Management Software"
> 	There are no examples in the website.

>From the name it could equally well be some piece of corporate
human-resources department software.

>       If it is things like
> 	hdparm, hwtools, powertweak, systune, we currently have them in
> 	admin::configuring; we can think of
> 	admin::configuring:performance maybe?

I'd say this fits under admin::hardware, although we might want to
split hardware setup from hardware monitoring.  More urgent is
better coverage under hardware::.

> "Problem Management Software"
> 	We miss the category.  We do have devel::bugtracker, but we do
> 	not have anything for trouble ticket trackers for sysadmins.
> 	Something for packages like otrs or gnats.
> 	PROPOSED: admin::trouble-ticket, although I feel like there can
> 	be better names.

admin::issuetracker?

> "Emulators"
> 	it can be mapped to hardware::emulation
> "Virtual User Interface Software"
> "Virtual Machine Software"
> 	We indeed miss categories for virtualization tools.
> 	PROPOSED:hardware::virtualization? Or admin::virtualization?
> "Software Configuration Management Tools"
> 	There are no examples in the website.  Can it be mapped to
> 	admin::configuring as well?

An admin:: version of the devel::versioning tag?

With this and the last, if a tag is turning up in more than one of
the "specialist use" facets, maybe it should be given a generic
use:: tag.  Though not unless some examples turn up.
 
> "Network Monitoring"
> 	iterating.org proposed to map to "security::ids"; it can also be
> 	mapped to admin::monitoring, use::monitor, network::scanner,
> 	depending on cases.  Do we need a network::monitoring?

Well, where does bandwidth-monitoring fit?  That's the one I
remember having trouble tagging.
 
> "Network Security"
> 	iterating.org proposed to map to security::antivirus and
> 	network::firewall.  We also have the whole security::* set of
> 	tags.
> 
> "Enterprise Connectivity Software"
> 	There are no examples in the website and I don't understand what
> 	this means.  We do not have much related to "Enterprise" things,
> 	so I'm interested.

When I ask google I find myself reading pages where people are
apparently offering to sell me a telnet server (maybe even some kind
of fancy modern SSL-enabled one) for several hundred dollars.

> "Remote Control Software"
> 	The best we have is I think use::login.  Can we do better?

Is this for things like ssh, krdc, stonith, or what?  They should
probably be use::login + network::client or something similar.

>  * "Website & Communication Applications"
> 
> This is much better done at iterating.org than in our web::* facet.
> Let's see what we can pick:
> 
>   "Content Management","web::cms",,,
>   "Content Management Server","web::cms",,,
>   "Authoring Software","web::cms",,,
>     I don't know if we currently need the distinction
>   "Translation/Globalization","web::cms",,,
>     We do have pootle, which I wouldn't call a CMS.  I can see a new
>     category coming, although it may be too early now if we only have on
>     package.  It can go under web::application, and further refined when
>     more packages come around.
>   "Forums"
>     Iterating proposes the 'bbs' tag.  PROPOSAL: I think we really need
>     web::forum
>   "Portals"
>     PROPOSAL: iterating.org proposes web::portal.  It's good to me!

Ah, now these do sound like the kind of thing that belongs under
web::.

>   "PHP Scripts"
>     this maps to web::* and implemented-in::php
>   "Search Engines"
>     PROPOSAL: iterating.org proposes to add web::search-engine
>   "E-Commerce Software"
>     PROPOSAL: iterating.org proposes to add web::e-commerce
>   "Source & Site Protectors"
>     PROPOSAL: iterating.org proposes to add web::site-protectors
>     However, there are no examples in the website and I don't understand
>     what it means.

"SiteProtectorTM" seems to be an IDS...
 
>  * "Email, Messaging & Other Communications"
> 
>   "Integrated Collaborative Environments"
>     What do we have in Debian?

Google shows me this slogan being used for a variety of things, some
mail-based (eg MSexchange), some im-based, some wiki-like; if it's a
category at all, it sounds like a use::.

>   "Standalone Email Applications"
>     It can be mapped to mail::user-agent
>   "Instant Messaging Applications"
>     It can be mapped to works-with::im
>   "Unified Messaging Applications"
>     This is something I do not know

Again it looks more like a trademark than a term anybody should
expect to be able to search on. 

>   "Conferencing Applications"
>     "apt-cache search conference" shows various results, so we need a
>     tag.  Problem: I don't know in which facet we should file it.

It makes me wonder if we should have had a comms:: (?) facet instead
of mail::, with tags for smtp, nntp, hamradio, irc, im, voip etc.
Probably too late now.
 
>  * "Security"
>   "Threat Management Software"
>     PROPOSAL: iterating.org proposes security::threat-management
>     There are no examples in the website, though, so I don't understand
>     this very well.

I'd be interested to hear of a package that belongs under security::
and *isn't* about managing security "threats" - usually with the
intent of modifying them in a generally downward direction.

>   "Secure Content Management"
>     PROPOSAL: iterating.org proposes security::content-management
>     There are no examples in the website, though, so I don't understand
>     this very well.  Would this be something like secure storage, for
>     packages like cryptsetup?

No idea.  Security through obscurity!

>   "Security and Vulnerability Management Software"
>     PROPOSAL: iterating.org proposes security::vulnerability-management
>     There are no examples in the website.  Is it for things like nagios?

Another one with low distinguishing power.  If there's no
"Vulnerability", there's nothing for your security software to do,
and all the word "Management" contributes is the unwanted
implication that you're not trying to eliminate root exploits so
much as making sure they're all properly catalogued.

>   "Adware & Spyware Removal"
>     PROPOSAL: iterating.org proposes something like
>     "security::malware-removal".  We can probably turn
>     security::antivirus into something that includes a broader concept
>     of malware, although AFAIK we do not currently have anything
>     specifically targeting adware and spyware in Debian.

Other than debian-installer...

>   "Online Privacy"
>     PROPOSAL: iterating.org proposes security::online-privacy.
>     security::privacy sounds even better.  I'm thinking of packages like
>     anon-proxy, libapache2-mod-removeip, privoxy, premail.

This makes sense.

>   "Parental Filters"
>     PROPOSAL: iterating.org proposes security::parental-filter.
>     I don't know if we have anything like that in Debian, and I
>     don't like the idea of parental filters at all, so if we need this
>     tag it won't be me that commits it to the vocabulary.

We have things that can function as content-censoring filters, but
nothing that comes pre-equipped with a canonical list of naughty
words.  Concerned parents wouldn't want such files to be available
for download anyway, should they? 

>   "Pop-Up Blockers"
>     PROPOSAL: iterating.org proposes security::pop-up-blocker.
>     I don't know if it has to do with security.  It can probably group
>     together ad blockers like bfilter and filterproxy.  Should we do a
>     "web::content-filter" and in that way we have a place to fit
>     parental filters (blargh) as well?

Filtering is also something mailsystems do, not to mention a lot of
random commandline utilities - should this be factored out as a
use::filtering? 

>   "Biometrics"
>     PROPOSAL: iterating.org proposes security::biometrics.
>     Do we have anything in Debian?

Not detectably.

>  * "Multimedia & Graphics"
>   "Audio Libraries Management"
>     PROPOSAL: iterating.org proposes sound::library-management.
>     I would probably prefer to use together works-with::audio and
>     use::organizing.

Agreed.

>   "Video"
>     PROPOSAL: iterating.org proposes a video facet, just like we have a
>     sound facet.  It makes sense.  Proposed tags are:
>     "Video Files Editing" -> "video::editors"
>     "Video Players" -> "video::players"

No, these are use::s.

>     "Video Codecs" -> "video::codes"

That's role::shared-lib (cf "The Theora Video Compression Codec", ie
libtheora0).

>   "Graphic"
>     PROPOSAL: iterating.org proposes a graphics facet, just like we
>     have a sound facet.  It makes sense, to me, only if it'll contain
>     tags that go beyond "works-with::image plus some of the use::*
>     tags", and can include something specifically related to the
>     professional world of image processing.
>     The current proposals would already have mappings:
>       "Image Editors" -> "graphics::editors"
>         can be done with "works-with::image, use::editing"
>       "Image Viewers",,,,"graphics::viewers"
>         can be done with "works-with::image, use::viewing"
>       "Graphics Conversion",,,,"graphics::conversion"
>         can be done with "works-with::image, use::converting"
>     There would be 3D-related proposals:
>       "Modeling and Animation" -> "graphics::3D"
>       "3D Modeling Software" -> "graphics::3D::modeling"
>       "3D Animation Software" -> "graphics::3D::animation"
>       "3D Rendering" -> "graphics::3D::rendering"
>     But then animation would fit into video as well.  I'm a bit
>     confused, and I'm wondering if we shouldn't pack everything
>     (including the current sound::* facet) together into some
>     multimedia::* facet.

I was in favour of replacing "Section: sound" with "Section:
multimedia" before I ever heard of debtags, so you'll hear no
arguments from me against (multi)media::audio, ::video, ::image,
::3d... other than arguments against then adding duplicated
media::*:editing tags, that is. 
 
>  * "Desktop Environment & Drivers"
>    I'm not sure why bringing the two things together.

Some kind of LiveCD association?

> "File Management" -> "admin::file-distribution ?"
>   You probably want use::browsing and works-with::file
> "Desktop Environment Software"
>   -> x11::display-manager
>   -> x11::font
>   -> x11::screensaver
>   -> x11::theme
>   Yup.  This is very broad and instead of a broad tag we try to have
>   tags to describe the specific functions of packages.
> 
>  * "Office & Business"
>    PROPOSAL: iterating.org proposes a new office facet, with these tags:
>    "Spreadsheet Applications" -> office::spreadsheet
>    "Scheduling Applications" -> office::schedule
>    "Front-Ends" -> office::front-end

(There's a package of debian-front-desk?  Or what would it be?)

>    "Office Suites" -> office:suites

s/suites/suite/

>    "Office Add-Ins" -> office:add-ins

What would this be?  (Freecell?)

>    "Sales & Marketing Tools" -> office:sales-marketing
>    "Business Finance" -> office:business-finance
>    "Document Management" -> office:document-management
>    "Legal" -> office:legal
>    "Personal Finance" -> office:personal-finance
>    "Personal Info Managers" -> office:info-mangers
>    "Presentation Tools" -> office:presentation
>    "Project Management" -> office:project-management
>    That's much more insight of software used in real offices that we
>    ever had in debtags.  Maybe an office facet is really to be
>    considered to support the point of view of that kind of environment;
>    I'm not a field expert here, but if someone who does wants to
>    maintain such a facet, I'd like the idea.

Generally speaking this does sound good.

>  * "Storage"
>    "Archive and HSM" -> admin::backup
>      you can also use works-with::archive
>    "Backup and Recovery" -> admin::forensics
>      it's probably still admin::backup rather than forensics

I have in the past thought about suggesting that we needed an
admin::recovery (or ::rescue) more than an admin::forensics;
"forensics" tends to be a sort of recovery + security combo.

>    There's a proposal for a new storage::* facet:
>      "Storage Replication Software" -> storage::replication
>      "Storage Resource Management Software" -> storage::management
>      "Capacity, Quota, and Resource Management" -> storage::management::capacity-quota
>      "Device and SAN Management" -> storage::management::device-SAN
>      "Storage Policy and Automation Software" -> storage::management::automation-policy
>      "Virtualization and Volume Management" -> storage::management::virtualization-volume
>    I'm no field expert here either, and I don't know if we can instead
>    do with combining the existing facets we have.

Well, we have an overworked hardware::storage, but I'm not convinced
these would help there.  Ideally admin::filesystems would include
things like disk-partitioning setups too (and therefore mdadm, EVMS
etc as well as e2fsprogs), but the name doesn't quite fit and I
can't think of any clear improvement.
-- 
JBR
Ankh kak! (Ancient Egyptian blessing)