[IAEP] SoaS as a Sugar Labs project.

Gary C Martin gary at garycmartin.com
Tue Aug 25 21:22:40 EDT 2009

On 25 Aug 2009, at 16:49, Gary C Martin wrote:

> On 25 Aug 2009, at 16:03, Michael Stone wrote:
>>> El Mon, 24-08-2009 a las 20:58 +0200, Martin Langhoff escribi=F3:
>>>> And also... and completely from the outside... I'll apologise in
>>>> advance for saying something I know might be controversial. I worry
>>>> that SL seems to have -- for a external party like me -- more
>>>> bureaucracy than it has people "doing". IMHExperience, the
>>>> projects I
>>>> enjoy working on, and that I see being productive have  a much  
>>>> lower
>>>> "procedure/label/committe " : contributor ratio.
>>> I don't necessarily disagree with you, but just 2 days ago I was
>>> offered
>>> an advice on the other side of the spectrum by Michael: he notes
>>> that a
>>> lot of important things are falling through the cracks because  
>>> nobody
>>> organizes the available resources.  His suggestion is to introduce
>>> real
>>> project management into the game, which is basically what David's
>>> Projects idea seems to bring.
>> For the record, I consider my puny efforts to offer more support for
>> Martin's
>> and Greg's remarks than for David's.
>> (The analysis is simply that our current situation is unsurprising
>> given the
>> facts that, first, SL seems to consist more of leaders than of
>> followers and
>> second, that there seems to be a real dearth of people who care more
>> about
>> getting other people unstuck than about making progress on their own
>> pet
>> projects.)
>> (Though, obviously, I'm more guilty than most here.)
>>> A meta-comment on your post: you don't need to apologize and be shy
>>> for
>>> offering your criticism, no matter how many people will disagree  
>>> with
>>> you.
>> Actually, he does need to apologize and to be shy because doing so
>> makes it
>> easier for folks to hear what he's trying to say.
>> (In our current environment, it works rather similarly to good-cop/
>> bad-cop.)
>>> I recently got useful criticism from Bemasc, Christoph and Daniel on
>>> #sugar regarding our relationship with Deployments.  Their feeling  
>>> is
>>> that we didn't do enough to get them involved, mine is that our
>>> efforts
>>> to reach out have been largely unsuccessful for reasons I do not
>>> fully
>>> understand.
>> Here's another reason for you to consider:
>> I have come to believe that many people involved in deployments have
>> *learned*
>> that they're not going to get anything useful out of interacting
>> with SL
>> because:
>> 1. SL has largely ignored the feedback supplied by these deployments
>> in
>> 2007-2008 and exhaustively documented by Greg Smith and S Page at
>>  http://wiki.laptop.org/go/Feature_roadmap#Roadmap
> Wow, blast from the past :-) Actually I'd strongly disagree here.
> Having re-read through most of what is listed here, much progress has
> been made on a large number (dare I say majority) of these items!

OK, I took up my own challenge here...

So, please do treat comments below as from the peanut-gallery, and not  
as authoritative responses to these ~82 feature requests – many of  
them significant compound feature sets in of themselves. I'd say close  
to 50% have been resolved, 10% are specifically XO hardware related,  
10% XS server, 15% I don't know enough to confirm or deny, 5-10% are  

FWIW: There's a cold beer waiting at the end of this email.

-- Feature roadmap/Spell checker in Write:

English checking is back in and working, though unsure about other  
languages, likely just down to availability of a dictionary file for a  
given language.

-- Feature roadmap/Sugarized color picker:

Done and in working for Write, Paint Activity (and others that use  
colour) haven’t picked this up yet, just a matter of free time (I  
almost did Paint but keep getting distracted).

-- Feature roadmap/Easy "Sugarization":

How long is this piece of wet string ;-) But there has been lot’s of  
effort here. Off the top of my head I’d point to the switch to  
Metacity in 0.86, Tomeu’s GTK+ widget for making Gnash content into  
full Activities, Tomeu’s hellow-world PyQt based Activity example.  
Benjamin’s GSOC Groupthink workhttp://bemasc.net/~bens/groupthink/groupthink-module.html 
   for solving collaboration simply (for the author) in many classes  
of Activity type. Lucian’s GSOC Webified work http://honeyweb.wordpress.com/ 
  for creating custom site specific browser Activities. Universal  
bundle work from Aleksey came close but pused to 0.88 for time/ 
stability reasons.

-- Feature roadmap/Activity updater improvements:

This was part of OLPC distro onlu for a long time, but is now recently  
over as a part of Sugar. Majority of work I’m aware of is to get it  
working smoothly with the activities.sugarlabs.org mozilla addons  
based site (a big improvement from all the manual wiki hacking).

-- Feature roadmap/Concept maps:

Me, me, ME!! ;-) No, actually Tomeu picked up Labyrinth sugarisation,  
then I did some UI polishing and hooked up extra features along with  
the talented Aleksey. http://wiki.sugarlabs.org/go/Activities/Labyrinth

-- Feature roadmap/New activities:

Mixed list here. Abiword has always allow pasting of images to my  
knowledge. Socialcalc is still rolling along with updates from time to  
time. The switch to Metacity for 0.86 allows any application to be run  
if you really want to go that way (and if the HW can actually cope).  
Tomeu has added a Gnash GTK+ widget and an example PyQt hello world.

-- Feature roadmap/Browse update and multi-media:

Hmmm... did this get done? Yep, think it did.

-- Feature roadmap/Terminal improvements:

Hmmm... Somewhat vague requests. We’ve since got multi tab support  

-- Feature roadmap/Better eBook reader:

Lot’s of effort, new Activity options (James) and format support  
(Sayamindu) now available. Probably needs re-evaluation based on where  
we are now.

-- Feature roadmap/Single sign-on from Browse:

Pass. I’ve seen work happening, but having never tested or tried this  
side of school server work so don’t know where we stand today.

-- Feature roadmap/"Candy Bag" or "Bulletin Board" activity:

Aleksey created a very feature rich Library Activity - perhaps too  
feature rich ;-) – some of which is going into 0.86 Journal and is  
related to other proposed features now pushed to 0.88.

-- Feature roadmap/Scalable link-local presence:

Another long piece of wet string, but specifically Cerebro seems to  
have no one recently pushing on it for adoption. So most efforts are  
in debugging existing stack, picking up general fixes and features  
from Collabora, (scalability work seems aimed at server side).

-- Feature roadmap/Synchronous collaboration:

Big bag of different feature here. Some done (Journal ‘send to ->  
friend’ file sharing via both gabble & salut), some not, some now for  
0.88 (Gadget/JID stuff), some partial (buddy tagging is in same ball  
game, but more viable/useful vs. buddy avatars). Should also mention  
Benjamin’s GSOC Groupthink work solving collaboration for many classes  
of Activity type.

-- Feature roadmap/Collaboration groups:

Sure Martin mentioned this is implemented (using Moodle) but not  
necessarily tested/used much yet. Tomeu also implemented buddy tagging  
(aka profiles) – Gabble only at the moment, and the UI hasn’t made 0.86.

-- Feature roadmap/Asynchronous collaboration:

Long list of (server based) features here that seem to be a  
description of what I understand Moodle (or some other CMS) is  
designed to do... Some feature work for Browse in 0.86 should smooth  
some rough edges (upload/download file naming).

-- Feature roadmap/File sharing:

For user initiated transfers Tomeu’s Journal ‘send to -> friend’  
feature solves much of this. Aleksey’s Library Activity is aimed at  
the ad-hox ‘making some Journal stuff generally available’. Solutions  
for getting the latter into 0.86 Journal were considered, but pushed  
to 0.88.

-- Feature roadmap/Journal improvements:

Some of this is done (Tomeu’s Datastore rework, specifically for  
robustness). But plans have moved on given the SL volunteers able &  
willing to do the work.

-- Feature roadmap/Hardware alerts:

The specific use case (95% disk space) was already covered (though a  
little crudely). Notifications are running quite well but don’t cover  
devices yet (I think), which would likely be the correct place for  
these alerts.

-- Feature roadmap/Contextual help:

Not much movement here yet, at least we have more accelerator keys  
getting shown in palettes so you can learn the keyboard shortcuts in  
context :-) (FWIW, my vote is to focus on making Activities *simpler*  
so they do not need help text bloat, and use better secondary help  
hint messages).

-- Feature roadmap/GUI suggestion from SJ:

This is a path to walk rather than a feature. Think we are making good  
progress down this path.

-- Feature roadmap/Trash can:

No description or specification at all on this one. I’m guessing  
“Trashcan for Journal” no work done I’m aware of.

-- Feature roadmap/Scalable zoom levels:

A Neighbourhood list view almost made an 0.86 appearance, but think  
it’s pushed to 0.88 now. Same goes for additional UI design work  
needed for Gadget. Though Tomeu’s buddy tagging is in 0.86 (but no UI  
exposure and Gabble only).

-- Feature roadmap/Keyboard navigability:

Many toolbar palettes now clearly display their accelerator keys, and  
several views are now using the GTK+ treeview for 0.86, so inherit  
better keyboard navigation accessibility.

-- [XO HW] Feature roadmap/Grab-scroll keys:

It is implemented in several Activities (Labyrinth uses it to pan the  
canvas), and there is an (XO hardware specific) keyboard shim type  
background daemon that has been demonstrated as a potential general  

-- [XO HW] Feature roadmap/Blinking lights usage

No description, specification at all on this one. Pretty sure Mikus  
was after the voodoo that drives the XO HW wireless leds.

-- [XO HW] Feature roadmap/VGA Projector:

Not much SL can do. Though there is a VNC screen sharing activity or  
two that may help, along with some better Activity collaboration.

-- [XO HW] Feature roadmap/Accurate touchpad:

No comment ;-) Actually Sugar can help with better keyboard navigation  
and key discovery (both have made good progress). FWIW Pail Fox has  
been reviewing and making more touch pad discoveries, there is a new  
driver in the F11-on-XO1 tree, but I don’t think it’s in the latest  
available build yet.

-- Feature roadmap/Run Fedora applications on XO:

0.86’s move to Metacity will help a lot here, but this feature is  
largely obsoleted now that the F11-on-XO1 OLPC builds ship both Sugar  
(the default) and GNOME desktop, and have a button to reboot between  
the two if desired.

-- Feature roadmap/Copying files between Journal and USB:

Done. Also Tomeu’s 0.84 datastore work made using a USB stick between  
different operating systems *much* more reliable (no ghosts, no  
vanishing files, no odd doppelgänger).

-- [XO HW] Feature roadmap/Replace JFFS file system with better one:

Only a low level FS issue linked to XO HW needs really, nothing SL can  
do about this. XO1.5 will avoid this issue due to HW changes.

-- [XO HW] Feature roadmap/NAND full should not crash XO:

Done (though somewhat crudely).

-- Feature roadmap/File name and directory access:

Tomeu’s 0.84 datastore work improves some of this, it is just a simple  
file directory structure. Could perhaps do with a trivial ls-journal  
(or some such name) script to make terminal Journal browsing better  
for novice use. The copy-to-journal and copy-from-journal already work  
well for most needs.

-- [XO HW] Feature roadmap/Rebase on Fedora 10:

Done, well obsolete. Ongoing F11 and F12 work. SoaS work has also been  
a help here getting in bug reports etc.

-- Feature roadmap/Spell checking:

Sayamindu has an interesting ‘dictionary everywhere’ effort http://people.sugarlabs.org/sayamindu/global_dictionary.ogv 
  Not sure why, but seems to have slipped 0.86. I could also see a  
smaller version as a possible ‘dictionary device’ palette in the frame.

-- Feature roadmap/Better Arabic Support:

Pass. Sorry don’t know enough about this to comment.

-- Feature roadmap/Enhanced i18n:

Don’t know the details, but I think quite a chunk of work has been  
done here (by Sayamindu). Also being able to select several fallback  
languages in 0.86 I’m sure is a big help.

-- Feature roadmap/SCIM:

Pass. Sorry don’t know enough about this to comment.

-- Feature roadmap/Language customization:

Think this is done as well but not enough detail in feature  
description to know.

-- Feature roadmap/Language pack version 3:

Pass. Sorry don’t know enough about this to comment, though I have  
seen a fair amount of discussion about issues it raises.

-- Feature roadmap/RTL support:

Duplicate feature of above ‘Better Arabic Support.’ Sorry don’t know  
enough about this to comment.

-- Feature roadmap/Multilanguage support:

Completed for 0.86 (a new CP lets you select languages in order of  

-- Feature roadmap/Translations:

Pass. Sorry don’t know enough about this to comment.

-- Feature roadmap/Network Manager GUI:

Pass. Sorry don’t know enough about this to comment.

-- Feature roadmap/Document and improve default network connection:

Well, Sugar is much better auto-connecting quickly,correctly to an AP  
as the whole Mesh hunt run around is not currently supported. But,  
sorry, not sure how much else listed here has been tweaked.

-- Feature roadmap/802.1x support:


-- Feature roadmap/Full IPv6 support:

Pass. Maybe we picked up improvements due to F11 and F12?

-- Feature roadmap/XO as internet gateway (formerly called MPP):

Hmmmm. I seem to remember reading the Tomeu’s new ad-hoc 0.86 support  
allows internet sharing, but have not discovered this myself yet  
during testing, so likely ignore me. I understand MPP can be enabled  
manually, but is disabled by default due to wireless bandwidth  
saturation issues.

-- Feature roadmap/"Asynchronous internet":

Not much progress, Lucian’s GSOC Webified work may help a little here.

-- Feature roadmap/Reliable access to encrypted APs:

Better I think, but likely not ‘the best’. The joys of working with  
poor wireless security specs.

-- Feature roadmap/Printing support:

Vamsi’s GSOC Printer support covers this I think http://wiki.sugarlabs.org/go/Print_Support 
  though not sure when it will land (USB and server support via Moodle).

-- Feature roadmap/Debug tool:

Hmmm... “Diagnostic USB stick which analyzes and reports back detailed  
status.” A some what open ended feature request. Pass. Sorry don’t  
know enough about this, though Sebastian has been pushing on smolt for  
semi-automated hardware feedback – but I guess this feature was a  
request for some ‘Sugar’ or ‘Fedora’ failure case detection?

-- Feature roadmap/Backup to Internet:

Pass. Sorry don’t know enough to comment on status, though it does  
seem to be actively worked on (SoaS backup to a school server seems  
particularly active topic).

-- Feature roadmap/Universal view-source:

Done, perhaps some custom binary file types the source viewer doesn’t  
display yet, but source & syntax high-lighting is all looking lovely  
and shiny. Discussion moved on a long time ago (almost a year?) to  
source code search, universal edit-source and versioning. Along with  
source view for Sugar code its self.

-- Feature roadmap/Developers Center web site:

Done. See Sugar Labs wiki and infrastructure.

-- Feature roadmap/System clock:

“Must have a clock on the XO” I’ve been running Martin Dengler’s frame  
device clock for a while. It’s a bit visuallu large (i.e. device non- 
square icon) and doesn’t have an analogue mode, but any deployment can  
choose to add it to their builds, no questions asked, it's written as  
a Sugar 'extension'.

-- Feature roadmap/Library and bundle updates:

Again... Long piece of wet string topic. Lot’s of discussion, Aleksey  
has some great universal bundle proposals, but they’ve now been pushed  
back to 0.88 due to lack of support and worries of stability (due to  
development/QA time).

-- Feature roadmap/Caps lock option:

Lack of a Caps lock key is a major feature, but lack of a global Caps  
lock option (uppercase display of all text) is a missing feature.  
Walter hoped to pick this up for 0.86, but apologised for letting it  
slip to 0.88.

-- Feature roadmap/Fully comply with GPLv3:

Someone phone the lawyers! No seriously, report any missing legal asci  
and get it actioned. Developers know it’s (annoyingly) important even  
if fixing bugs, making things work, and adding new features are  
foremost in our minds.

-- Feature roadmap/Screen zoom:

No specific progress. But with Metacity in 0.86, I’m sure you can run  
any old screen zooming software you care too. Smart, clever Sugar UI  
integration will need to be for another day.

-- Feature roadmap/General UI sluggishness:

Again, more wet string but a laudable goal. For non-XO HW I believe  
compositing is now enabled by default. Metacity is better supported  
but I’ve seen no ‘better/worse’ response measurements for it (though  
allegedly it does eat more memory than Matchbox did, so that’s not a  
great sign). Lot’s of bug fixes and design  clean-up, but it’d take  
some focused QA time to see if there are real measurable UI changes on  
average, and you need to argue that against the improved feature set.  
FWIW, I at least haven’t noticed a degradation in XO performance. But  
this is a ‘needs testing’ item.

-- Feature roadmap/Faster task switching:

Duplicate. See above.

-- Feature roadmap/Faster activity launch and save:

Duplicate. See above.

-- Feature roadmap/Shutdown menu:

Paul Foxes powerd see current F11-on-XO1 implements this, best I can  

-- [XO HW] Feature roadmap/No power regressions:

OK, so not just XO HW, but that’s the intended regression case here.  
Paul Fox has done some wonderful stuff in powerd (XO, only is my  
understanding), seemingly much more than in previous OLPC builds, you  
can see it in action on F11-on-XO1. Though there is still active  
discussion if OLPC will be shipping it in their builds.

-- [XO HW] Feature roadmap/Improved battery life:

Again not just XO HW but that is the assumption given Sugar was not  
reliably available elsewhere at the time. Duplicate, see above. Though  
this does include a long shopping list touching just about every part  
of HW, device driver level, screen, network, etc, etc you could  
possibly try to mention.

-- Feature roadmap/Clipboard:

Hmmm, all information about this appears to have been purged from the  
wiki. Must have been real good stuff ;-) FWIW: The Clipboard does now  
show a preview of text you clipped making life easier if you have a  
bunch of clippings in use.

-- Feature roadmap/Memory pressure:

Both 0.84 and 0.86 have introduced UI changes to improve reported  
causes of memory pressure. The former implemented a home view that  
resumes by default and so aid as a switch to current instances rather  
than create new instances one after another. The latter implements the  
new tool bar design that had a primary goal of always keeping the  
“Stop” toolbar button visible so that folks didn’t get lost and could  
stop existing activities clearly at all times. Though still no low  
level resolution, once you stray into the woods Fangorn.

-- Feature roadmap/Journal never loses work:

Activities improve all the time... And Tomeu did a top notch job on  
the robust Journal re-implementation in 0.84. Unfortunately most  
deployments still don’t have it and failures are from ~year + old  
software :-(

-- XS Blueprints:Datastore Simple Backup and Restore:

I haven’t seen this in action, but I understand this is available and  
has been for some time (Martin can you comment?).

-- Feature roadmap/Image customization:

SoaS (distro) has re-raised many of these cases (random cloning of  
already booted and hand hacked sticks has caused much QA traffic).  
Many point seem more pertinent a question for the base OS distros, and  
not Sugar, which seems pretty flexible in how it is customisable.

-- [XO HW] Feature roadmap/Image signing key delegation:

Done I think. And future XO-1.5 HW is unsigned by default as I  
understand it.

-- Feature roadmap/Activation lease security:

Pass. Sorry don’t know enough to comment on status, though Martin seem  
to be actively working on such stuff.

-- Feature roadmap/Theft reporting:

Duplicate. Pass, same as above.

-- Feature roadmap/Activation via wireless:

Pass. Sorry don’t know enough to comment on status, though I do  
remember reading about many such items being worked on (Martin again I  

-- Feature roadmap/XO monitoring:

Pass. Sorry don’t know enough to comment on status, some of this may  
be possible now via the XS and client strings reported by Browse  
(linked to activity.SL.org and XS registration processes). Also plenty  
of discussion regarding the benefits and drawback of ‘assessment’  
based on stat collection.

-- Feature roadmap/GUI OS updates:

Pass. Sorry don’t know enough to comment on status, though there is  
ongoing work now that the update control panel had been picked up as  
part of Sugar – mainly related to supporting the  
activities.sugarlabs.org infrastructure. Full OS updates still seem a  
fairly open question.

-- [XO HW] Feature roadmap/Faster imaging:

For the XO, can I just say Nand Blaster, lightning fast, robust, and  
dead simple to use! All hail The Mitch!

-- Feature roadmap/Improved antitheft mechanisms:

Pass. Sorry don’t know enough to comment on status. Likely see Martin  
re: above as well as he’s been working here recently.

-- Feature roadmap/Security and isolation work:

Pass. Embarrassed to say I don’t know enough to comment on status.  
Rainbow had fallen out of Sugar for a while, but I understand much of  
the underlying work is now back in place. This is currently more an  
area of regression than improvement.

-- Feature roadmap/Scalable server-based presence:

Long piece of wet string, again, work progresses... SoaS team have  
recently been pushing on the server side to try and establish the weak  
points. Gadget is implemented, but as yet no GUI for it on the Sugar  

-- Feature roadmap/School server push of XO images:

Pass. Sorry don’t know enough to comment on status.

-- Feature roadmap/Automatically recognize XO on restore page:

Pass. Sorry don’t know enough to comment on status.


P.S. Wow, you made it down the page! I fibbed about the cold beer,  
sorry, I needed it myself about two and a half pages back.

More information about the IAEP mailing list