Discussion:
The future of OFLB development
Daniel Johnson
2012-12-03 15:31:26 UTC
Permalink
Hi all,

Dave Crossland's recent email reply has indicated that OFLB (and Aiki)
is no longer being actively developed, and he has suggested Django or
Drupal as a basis for future development. I am not only an amateur font
designer but also a senior Drupal developer. I've led teams in
development of dot-gov Drupal sites as well as in the private sector.
I'd be willing to take the development lead in getting this project off
the ground if there's sufficient community interest and support. Any
takers?

Cheers,
Daniel
Dave Crossland
2012-12-03 15:40:51 UTC
Permalink
AWESOME :)

Yes please
Garrick van Buren
2012-12-03 15:47:59 UTC
Permalink
The original Kernest.com font directory was built on Ruby-on-Rails, not a direction that I'd do again. Lately, on both the personal and professional side, I've been exploring MediaWiki as an application platform, combined w/ the use of the SemanticMediaWiki extension - it provides a lot of capabilities (versioning, reverting, account model/permissons, discussion, omature and popular framework). Additionally, the interwiki links and ability to upload to Commons.Wikimedia.org feel like a strong opportunity for the persistence of OFLB content. The downside is that mediawikis theme-ing isn't as straightforward as other CMSs.

As an example of this kind of wiki, over the past year, I've been working on
http://wiki.planetkubb.com

Of specific interest to the robustness that MediaWiki supports, check out a Game.
http://wiki.planetkubb.com/wiki/US_Nationals_2012,_Finals,_Knockerheads_v._Kubbsicles,_July_15_(Game_2)

-----------------------
Garrick van Buren
612 325 9110
garrick-***@public.gmane.org
-----------------------
Kernest.com
eBooks. Typography. Together.
-----------------------
Post by Daniel Johnson
Hi all,
Dave Crossland's recent email reply has indicated that OFLB (and Aiki) is no longer being actively developed, and he has suggested Django or Drupal as a basis for future development. I am not only an amateur font designer but also a senior Drupal developer. I've led teams in development of dot-gov Drupal sites as well as in the private sector. I'd be willing to take the development lead in getting this project off the ground if there's sufficient community interest and support. Any takers?
Cheers,
Daniel
Dave Crossland
2012-12-03 16:04:59 UTC
Permalink
I like the idea, but Drupal has a huge developer community....
Garrick van Buren
2012-12-03 18:27:11 UTC
Permalink
Lots of frameworks have a huge developer community. Here in MN, the WordPress and Rail communities rival, if not exceed, the local Drupal community. The satellite Microsoft office also gave rise to a vibrant developer community around their tools. In my experience, every framework has benefits and weaknesses.

The question to me is about long term appropriateness for this project.

Lastly, if the weaknesses - in the short term - of remaining in Aiki aren't detrimental to the OFLB. Then, there's time to discuss what will support the most successful OFLB.

-----------------------
Garrick van Buren
612 325 9110
garrick-***@public.gmane.org
-----------------------
Kernest.com
eBooks. Typography. Together.
-----------------------
Post by Dave Crossland
I like the idea, but Drupal has a huge developer community....
Dave Crossland
2012-12-03 18:58:11 UTC
Permalink
Post by Garrick van Buren
what will support the most successful OFLB.
The OFLB project has failed to attracted any developers beyond the
ones I raised funds for (most of which was out of my own pocket)

I believe a libre font library project that has a concrete roadmap,
openly calls for developers, and uses a mainstream framework will
attract developers.
Daniel Johnson
2012-12-25 16:27:54 UTC
Permalink
Drupal Pluses:
1. Good user management out of the box
2. Good theming separation-of-concerns
3. Well-maintained core code
4. Modular structure which allows for custom expansion.

Drupal drawbacks:
1. Memory hog
2. Error-prone caching system
3. Difficult to tune when massively scaling
4. Procedural coding style
5. Would require many hundreds of LOC to wrap the hierarchical data structures. Such custom code does not benefit from the many-eyes-shallow-bugs ecosystem of Drupal core.

As I see it, we have the following basic data structure:

Font contains metadata and one or more Versions.
Version contains metadata and one or more Faces (bold, italic, etc).
Face contains metadata.

I also think Alexandre's question is a very important one that needs an answer before we proceed in any direction. The one advantage OFLB has is that typeface authors become the curators of their own fonts, which we don't see so much on, say, Fontsquirrel, Kernest, or Google.

Sent from my iPhone
Post by Dave Crossland
I like the idea, but Drupal has a huge developer community....
Dave Crossland
2012-12-18 12:15:24 UTC
Permalink
Post by Garrick van Buren
The original Kernest.com font directory was built on Ruby-on-Rails, not a direction that I'd do again. Lately, on both the personal and professional side, I've been exploring MediaWiki as an application platform, combined w/ the use of the SemanticMediaWiki extension - it provides a lot of capabilities (versioning, reverting, account model/permissons, discussion, omature and popular framework). Additionally, the interwiki links and ability to upload to Commons.Wikimedia.org feel like a strong opportunity for the persistence of OFLB content. The downside is that mediawikis theme-ing isn't as straightforward as other CMSs.
As an example of this kind of wiki, over the past year, I've been working on
http://wiki.planetkubb.com
Of specific interest to the robustness that MediaWiki supports, check out a Game.
http://wiki.planetkubb.com/wiki/US_Nationals_2012,_Finals,_Knockerheads_v._Kubbsicles,_July_15_(Game_2)
What do you need to move forward with this? :)
Dave Crossland
2012-12-18 12:14:51 UTC
Permalink
Hi!
Post by Daniel Johnson
Dave Crossland's recent email reply has indicated that OFLB (and Aiki) is no
longer being actively developed, and he has suggested Django or Drupal as a
basis for future development. I am not only an amateur font designer but
also a senior Drupal developer. I've led teams in development of dot-gov
Drupal sites as well as in the private sector. I'd be willing to take the
development lead in getting this project off the ground if there's
sufficient community interest and support. Any takers?
What do you need to move forward with this? :)

Cheers
Dave
Daniel Johnson
2012-12-25 15:55:26 UTC
Permalink
Having looked at the required data model more in depth, and looking at my own feature wish list, I'm no longer convinced that Drupal is the right tool for this job. (Drupal doesn't handle hierarchical data models very elegantly.) I'd be curious to see what MediaWiki could do. From my own experience, this seems like something that could be best done in RoR or Symfony2, but I have no MediaWiki experience.

Regardless of the platform, I'd like to remain involved, whichever direction we go.

Cheers,
Daniel
Post by Dave Crossland
Hi!
Post by Daniel Johnson
Dave Crossland's recent email reply has indicated that OFLB (and Aiki) is no
longer being actively developed, and he has suggested Django or Drupal as a
basis for future development. I am not only an amateur font designer but
also a senior Drupal developer. I've led teams in development of dot-gov
Drupal sites as well as in the private sector. I'd be willing to take the
development lead in getting this project off the ground if there's
sufficient community interest and support. Any takers?
What do you need to move forward with this? :)
Cheers
Dave
Alexandre Prokoudine
2012-12-25 16:09:06 UTC
Permalink
I'd start with a better question: what place in the modern ecosystem
should OFLB be aiming at?

When we started it, there was no TypeKit, no GFS, no half a dozen
other web fonts foundries.

What makes OFLB special today, apart from free-as-in-speech typefaces?

Alexandre Prokoudine
http://libregraphicsworld.org

On Tue, Dec 25, 2012 at 7:55 PM, Daniel Johnson
Post by Daniel Johnson
Having looked at the required data model more in depth, and looking at my own feature wish list, I'm no longer convinced that Drupal is the right tool for this job. (Drupal doesn't handle hierarchical data models very elegantly.) I'd be curious to see what MediaWiki could do. From my own experience, this seems like something that could be best done in RoR or Symfony2, but I have no MediaWiki experience.
Regardless of the platform, I'd like to remain involved, whichever direction we go.
Dave Crossland
2012-12-25 20:26:05 UTC
Permalink
On 25 December 2012 16:09, Alexandre Prokoudine
Post by Daniel Johnson
Post by Alexandre Prokoudine
I'd start with a better question: what place in the modern ecosystem
should OFLB be aiming at?
When we started it, there was no TypeKit, no GFS, no half a dozen
other web fonts foundries.
What makes OFLB special today, apart from free-as-in-speech typefaces?
I also think Alexandre's question is a very important one that
needs an answer before we proceed in any direction. The one
advantage OFLB has is that typeface authors become the
curators of their own fonts, which we don't see so much on,
say, Fontsquirrel, Kernest, or Google.
I think Daniel is correct.

What sets OFLB apart from FontSquirrel and GWF is that Ethan and I are
gatekeepers for those services, deciding what is uploaded to each,
whereas OFLB is 'self service.'

I think it would make sense to add a 'link' object that points to
projects that are hosted and developed elsewhere (DejaVu, Libertine,
etc etc) so that OFLB really IS a 'library' - a complete index of all
libre fonts on the web that presents the fonts in a way that is
pleasant to browse.

I think the features of www.openhatch.org are also good, in that they
help people to become involved. OFLB can help type designers to set
out a project's roadmap and invite people to participate on particular
tasks.
Dave Crossland
2012-12-27 12:57:32 UTC
Permalink
Post by Dave Crossland
On 25 December 2012 16:09, Alexandre Prokoudine
Post by Daniel Johnson
Post by Alexandre Prokoudine
I'd start with a better question: what place in the modern ecosystem
should OFLB be aiming at?
When we started it, there was no TypeKit, no GFS, no half a dozen
other web fonts foundries.
What makes OFLB special today, apart from free-as-in-speech typefaces?
I also think Alexandre's question is a very important one that
needs an answer before we proceed in any direction. The one
advantage OFLB has is that typeface authors become the
curators of their own fonts, which we don't see so much on,
say, Fontsquirrel, Kernest, or Google.
I think Daniel is correct.
What sets OFLB apart from FontSquirrel and GWF is that Ethan and I are
gatekeepers for those services, deciding what is uploaded to each,
whereas OFLB is 'self service.'
I think it would make sense to add a 'link' object that points to
projects that are hosted and developed elsewhere (DejaVu, Libertine,
etc etc) so that OFLB really IS a 'library' - a complete index of all
libre fonts on the web that presents the fonts in a way that is
pleasant to browse.
I think the features of www.openhatch.org are also good, in that they
help people to become involved. OFLB can help type designers to set
out a project's roadmap and invite people to participate on particular
tasks.
Continuing to noodle on this, I suppose we (I) should reach out to the
users and ask them what they want out of OFLB.
Barry Schwartz
2012-12-27 20:39:55 UTC
Permalink
Post by Dave Crossland
Continuing to noodle on this, I suppose we (I) should reach out to the
users and ask them what they want out of OFLB.
Quit trying to be fancy and do so much at once. If you cannot provide
easy, editable uploads and easy, straightforward downloads with a
_usable_ font display in between, the rest is garbage.

Somehow OFLB has managed never to achieve any of this, or even come
close to it, while instead pursuing the next trendy framework or
advanced functionality, which is just as quickly abandoned.

Also you must never, ever, ever again delete everyone’s work and make
them re-participate from scratch.

Furthermore, uploads need to be managed, whether or not you like it,
or OFLB is a trojan horse site. Users need to be trusted people, or
submit only through a moderator. Likewise, wiki needs to be managed,
because we know already (based on experience) that OFLB will not get
the attention needed to deal with spam as it occurs.

In short: have extreme patience when it comes to doing anything other
than the basics -- don’t even think about it too much until you have
the basics firmly under control -- and never do anything badly.

If none of this seems attractive, an alternative is to move to other
projects. There was a need for a free software clipart site, but free
software fonts are easy to obtain already. Other sites may not have
the features one wants, but see what I said above about ‘the basics’
and ‘patience’.
Nathan Willis
2012-12-28 14:56:24 UTC
Permalink
Post by Dave Crossland
Post by Dave Crossland
On 25 December 2012 16:09, Alexandre Prokoudine
Post by Daniel Johnson
Post by Alexandre Prokoudine
I'd start with a better question: what place in the modern ecosystem
should OFLB be aiming at?
When we started it, there was no TypeKit, no GFS, no half a dozen
other web fonts foundries.
What makes OFLB special today, apart from free-as-in-speech typefaces?
I also think Alexandre's question is a very important one that
needs an answer before we proceed in any direction. The one
advantage OFLB has is that typeface authors become the
curators of their own fonts, which we don't see so much on,
say, Fontsquirrel, Kernest, or Google.
I think Daniel is correct.
What sets OFLB apart from FontSquirrel and GWF is that Ethan and I are
gatekeepers for those services, deciding what is uploaded to each,
whereas OFLB is 'self service.'
I think it would make sense to add a 'link' object that points to
projects that are hosted and developed elsewhere (DejaVu, Libertine,
etc etc) so that OFLB really IS a 'library' - a complete index of all
libre fonts on the web that presents the fonts in a way that is
pleasant to browse.
I think there's a definite need for a site that indexes all (that is, as
close to "all" as humanly possible) of the open fonts out there; it
definitely makes sense for OFLB to be that service. After all, the
foundries and webfont services like GWF are not going to be places that
point to fonts-that-live-elsewhere. That just doesn't make sense for them.
Like a library, being a resource that helps users find a font is
invaluable, whether the font is hosted at the site (the "local collection")
or somewhere else (interlibrary loan would be the analogue, but that's kind
of stretching the library metaphor to its limit).

That means having good and searchable metadata & samples (at least sample
images) for all of the indexed fonts hosted elsewhere, but on the other
hand it does not necessarily mean that OFLB needs to bother with *serving*
those fonts over HTTP. I mean, if GWF can deliver @font-face for a
particular font, then (a) OFLB doesn't gain anything by duplicating that
function for the same font -- and (b) it might even hurt the designer, if
Google were to ever implement any payment mechanisms.

On the other hand, there is clearly a need to be filled by serving up fonts
that are *not* hosted anywhere else. To me, the question here is how much
service makes sense to provide: just @font-face? also using OBS to build
installable packages? VCS?

If you look at OFLB from the graphic/web designer's POV, then providing
@font-face makes sense; we could do that for fonts that live only at OFLB,
and simply point to the correct HTML snippet for the fonts available at GWF
or wherever else.

The other things, like packaging or version control, are not as clear cut.
There are plenty of code hosting options for designers (from GitHub on up);
OFLB could provide pointers to them, which would help designers, but we
probably couldn't run a better one ourselves anyway. Packaging I'm on the
fence about -- I know there are a lot of independent font projects that
would benefit a lot by having packages available for Linux distros, but it
is kind of a different animal.
Post by Dave Crossland
I think the features of www.openhatch.org are also good, in that they
Post by Dave Crossland
help people to become involved. OFLB can help type designers to set
out a project's roadmap and invite people to participate on particular
tasks.
Continuing to noodle on this, I suppose we (I) should reach out to the
users and ask them what they want out of OFLB.
Agree. I think the "library/resource/index first" approach I described is
probably most useful to graphic designers & pub designers; they'd be the
ones to ask for feature input or feedback. I don't think I actually know
any, though....

Nate
--
nathan.p.willis
nwillis-eiP9NBaGPlk1WUs8F/Ki+***@public.gmane.org
aim/ym/gtalk:n8willis
identi.ca/n8
Peter Baker
2012-12-28 19:35:08 UTC
Permalink
I'd like to support the idea of OFL indexing Open Source fonts that
are hosted elsewhere. To speak as a frequent visitor to OFL and an
occasional font maker, it would be great to have a resource that
aggregates Libre fonts (or at least the major ones) and is actually
kept up to date.

Yet I can see major technical problems, especially if you intend to
include descriptions samples and the like. How would you pull fonts
from GitHub, SourceForge and elsewhere (each source presenting its own
set of problems, not to mention all the things released from the
designers' own webpages)? How do you cope with the fact that every
font is differently packaged (for there is no standard for packaging
fonts)? If the answer is to get font designers to cooperate, what are
the incentives? When I release a new version of my biggest font
project, I already have a lengthy checklist to get through: I'm not
eager to add the making and uploading of a new package for OFL, or
samples, or descriptions, to my list of stuff to do--just as I don't,
for example, think it my job to provide Linux packages or update the
Wikipedia page. If the answer is for the website's maintainers to find
and index the fonts, how do you ensure the continuity of the effort?

The idea of serving webfonts is I think a non-starter. The bandwidth
involved would be immense: who would provide the funding? And if
Google is providing the same service, what's the point?

Peter Baker
Nicolas Mailhot
2012-12-31 09:46:34 UTC
Permalink
Post by Alexandre Prokoudine
I'd start with a better question: what place in the modern ecosystem
should OFLB be aiming at?
When we started it, there was no TypeKit, no GFS, no half a dozen
other web fonts foundries.
What makes OFLB special today, apart from free-as-in-speech typefaces?
This apart is pretty important :)

It could be emphasised more by exposing all the font lifecycle (including
vcs, fontforge, system installation, etc) instead of focusing on the
read-only obfuscate-in-js-and-webfonts other web foundries promote. Right
now I'm not sure if a casual visitor will even realise those parts exist
--
Nicolas Mailhot
Nathan Willis
2013-02-11 20:33:37 UTC
Permalink
An addendum to this topic: there are evidently still software projects out
there whose users rely heavily on type and need a good font selection --
but aren't currently getting one.

Nicolas Spalinger just posted a link to this LibreOffice proposal:
https://wiki.documentfoundation.org/Design/Whiteboards/Font_Repository_Integration

...which is about integrating download & installation of open fonts into
LO. I at least think OFLB could fill this role better than anyone else.
Or at least OFLB _could_; I'm not sure exactly what their technical
requirements are at the moment....

Nate
--
nathan.p.willis
nwillis-eiP9NBaGPlk1WUs8F/Ki+***@public.gmane.org
aim/ym/gtalk:n8willis
identi.ca/n8
Dave Crossland
2012-12-25 16:18:11 UTC
Permalink
Post by Daniel Johnson
Having looked at the required data model more in depth, and looking at my own feature wish list, I'm no longer convinced that Drupal is the right tool for this job. (Drupal doesn't handle hierarchical data models very elegantly.) I'd be curious to see what MediaWiki could do. From my own experience, this seems like something that could be best done in RoR or Symfony2, but I have no MediaWiki experience.
Regardless of the platform, I'd like to remain involved, whichever direction we go.
Perhaps extending http://impallari.com/projects is our best bet, made
with http://kohanaframework.org/
vernon adams
2012-12-25 21:11:57 UTC
Permalink
I have been thinking for a while (have i not mentioned it?) That I thought oflb should have a bone fide web font server, where web authors can serve the fonts with a single line in the head tag a la GWF.
-v
Post by Dave Crossland
On 25 December 2012 16:09, Alexandre Prokoudine
Post by Daniel Johnson
Post by Alexandre Prokoudine
I'd start with a better question: what place in the modern ecosystem
should OFLB be aiming at?
When we started it, there was no TypeKit, no GFS, no half a dozen
other web fonts foundries.
What makes OFLB special today, apart from free-as-in-speech typefaces?
I also think Alexandre's question is a very important one that
needs an answer before we proceed in any direction. The one
advantage OFLB has is that typeface authors become the
curators of their own fonts, which we don't see so much on,
say, Fontsquirrel, Kernest, or Google.
I think Daniel is correct.
What sets OFLB apart from FontSquirrel and GWF is that Ethan and I are
gatekeepers for those services, deciding what is uploaded to each,
whereas OFLB is 'self service.'
I think it would make sense to add a 'link' object that points to
projects that are hosted and developed elsewhere (DejaVu, Libertine,
etc etc) so that OFLB really IS a 'library' - a complete index of all
libre fonts on the web that presents the fonts in a way that is
pleasant to browse.
I think the features of www.openhatch.org are also good, in that they
help people to become involved. OFLB can help type designers to set
out a project's roadmap and
Dave Crossland
2012-12-25 21:19:31 UTC
Permalink
Post by vernon adams
I have been thinking for a while (have i not mentioned it?) That I thought oflb should have a bone fide web font server, where web authors can serve the fonts with a single line in the head tag a la GWF.
I'm not sure this is very valuable. OFLB offers this today but without
the full set of formats, its not so much use. Also I'm not totally
sure our gratis hosting provider would want to foot a massive bill...
vernon adams
2012-12-25 21:31:32 UTC
Permalink
I can see that is an issue :-) but then you are looking only for future developments that do not generate serious bandwidth. I wonder what sort of bandwidth a full on oflb font server could generate?
Post by Dave Crossland
Post by vernon adams
I have been thinking for a while (have i not mentioned it?) That I thought oflb should have a bone fide web font server, where web authors can serve the fonts with a single line in the head tag a la GWF.
I'm not sure this is very valuable. OFLB offers this today but without
the full set of formats, its not so much use. Also I'm not totally
sure our gratis
Loading...