Personal tools
You are here: Home How to Help Universal Communications Where Nepal Notes

Notes

by Nepal Studies last modified Nov 21, 2008 07:26 PM

Random disorganized jottings to be incorporated elsewere.

Timeline

Display

Low power display based on OLPC technology used in XO-1 may be in mass production Q2 2009 (with preview samples for manufacturers Q1 2009) according to Pixel Qi. This might enable possibilities for intermediate devices that include decent screen as well as audio without being unacceptably power hungry.

Fedora 10 on XO

Release of Fedora 10 on XO is planned for 2008-11-25 for availability on (expensive) 4GB class 6 SD flash cards together with next G1G1 scheduled via Amazon at same time. Extensive QA being done, but some of it based on exploratory testing

. Expect substantial bug fixing after initial mass release to large numbers of G1G1 users. But once that is over we have a an excellent platform immediately available. Will also result in other good SD card XO distributions but Fedora is probably best base for a customized distribution anyway. Customization to XO still very helpful even if not needed on SD card. New version of Xfce at Xmas so update to that if Gnome or KDE too heavy.

Generation 1.5

7. Gen 1.5 technical specs are nearing completion, with expected updates including a faster LX processor, increased DRAM of the DDR2 variety, increased flash (sufficient to support dual-boot), a newer wireless chip, and rubber bumpers to improve screen protection, if possible. Target ship date for Gen 1.5 is Q1 2009. The target for Gen 2 is Q2 2010.

From OLPC News 2008-09-15

 

Logistics lead time of 3 months for firm order plus something like 1 month for shipment implies any actual deployment as temporary school servers would be generation 1.5. Assume it will have at least 512MB RAM instead of 256MB and at least 2GB flash (probably 4GB) instead of 1GB (unconfirmed). Improved wireless chip said to improve power consumption but that isn't relevant for temporary school servers. Increased flash also not very important since will still need external HDD anyway and can just add flash as required. More RAM and faster CPU always nice. Presumably implies more power consumption affecting OLPC but only positive for temporary school servers and OLPT. Presumably fixes for reliability problems eg with touchpad and antenna are already rolled in to current production runs. Target would be Gen 2 for One Laptop Per Child but conceivably Gen 1.5 for One Laptop Per Teacher. Temporary Gen 1.5 school servers will therefore still be optimal as OLPT laptops (and adequate for OLPC if that happens earlier).

CardPodCaster Features

Many of these may not be available in cheapest off the shelf models for initial order, but worth considering for future orders which could include both minor ECO (Engineering Change Orders) and major ODM (Original Design Manufacturer) design work.

SDHC

Check that card reader can handle SDHC, not just SD. Affects long term value when later wanting to use with high capacity flash cards.

USB2 and USB2GO and USB power

USB2 desirable for later use with high speed flash cards. USB2GO highly desirable for several possible firmware enhancements including copy between two CardPodCasters or from CardPodCaster to USB thumb drive, USB microphone/headset for higher quality audio input despite no line in, external keypad. Accept power in from USB enables use with larger external battery packs also used to charge speakers and mobile phones - without needing a DC input jack. However USB power output should still be available for copying to or from (or playing from) USB thumb drive and for powering USB microphone or headset.

Line In

Desirable unless availability of USB2GO and feasability of firmware upgrade to use for quality microphones and/or separate AM or FM broadcast receivers with external antenna confirmed.

External Antenna

Enables broadcast recording outside normal coverage area. Alternative is separate broadcast receiver to line in which may be better if connector adds significant cost or delay.

AM Receiver

Much wider coverage area (with lower quality audio) but requires (separate?) external antenna from FM receiver more easily available. Alternative is separate broadcast receiver (widely deployed already) to line in.

FM Receiver

Easily available, probably at negligible extra cost and useful for direct recording. But not essential given alternative via Line In.

Real Time Clock

Oddly omitted from easily available models. Should be negligible cost and highly desirable for timestamping recordings and automatic recording of broadcast programs (especially for access to channels available only for narrowcasts while most of their broadcast audience is asleep).

Firmware

For items below it should be possible to add firmware upgrades later, though any additional buttons may assist in simplifying user interface for such additions. Firmware source code availability is essential (either to be provided or perhaps by known chipset and boot/upgrade system enabling reverse engineering).

 

Lyrics, Text and Address Book

These (and others below) have all been done in www.s1mp3.org and similar project firmware mods. Not as useful as might be assumed due to difficulty reading with small LCD screen plus additional battery consumption for backlight. However could be of some use eg address book or text files (or even lyrics files) could supplement an audio version of market price data available via broadcasts and rural telecentres (click map) from www.agripricenepal.com

 

Could still be useful even without local fonts. But more useful if devangari (and any others needed) have also been done. Devangari should have already been done for Indian market.

 

A-B repeat

As well as direct use (especially when recording), the extra buttons add very little manufacturing cost but simplify user interface for other firmware upgrades.

 

Virtual Keyboard

Utterly horrible UI on small LCD screen. Prefer external keyboard via USB2GO. But does enable at least theoretical option for setting addresses and other options directly (eg through file names) instead of via recorded requests for manual registration entries to directory.

 

Costs

Temporary School Server (TSS)

Currently looks like XOs are available (for schoolchildren) qty 10,000 plus @ USD $210 each (ex works). Small addition for c.i.f. and significant addition for external USB HDD and DVD recorder. Main cost still power. Essential power for TSS is minimal if just used to copy flash cards. Actual use for distance education of teachers would be somewhere between the 5W for children and 20W for normal XS school servers (closer to 5W), for as many hours as teachers are able to use it at school. Solar feasible. Aim to combine with battery charging station supporting white led lighting to get to a level of say 400Wh per day, which could be generated in two hours per day with a tiny (300W) genset. This won't be obsoleted by subsequent arrival of grid or mini-grid and so has to be cheaper than solar almost anywhere. Also should be cheaper to generate locally once at that level than to transport batteries. (1000 x 30g / AA  2500mA is 30kg porter load for 2.5kWh). Detailed transport and charging costs needed.

 Technology

Much of the OLPC technology (especially Sugar) is primarily relevant to direct use by school children (and some still hasn't been built yet). We are just using the low power and low cost XO hardware with a more "normal" server and desktop environment. Sugar is now included in Fedora 10 so teacher's can familiarize with what kids will be getting later. Role as a TSS with just one per school makes many aspects largely irrelevant, including the excellent wireless mesh, presence and collaboration features. Some unique features that will eventually be useful for servers may or may not be adequately implemented in time for use in TSS design.

Bitfrost Security

Secure boot and recovery is critical for unattended servers also used as general purpose computers. Bitfrost may offer additional advantages. Unclear how much is usable and how much relies on unimplemented features of Journal.

Journal File System

Olpcfs

looks very useful in all sorts of ways, but not at all clear that it exists yet.

Deployment

See excellent OLPC deployment guide

. Lots of it can be removed as only need for full OLPC. But remainder should be taken as basis for a draft deployment guide to modify for TSS and other aspects. Also show the extent to which completing OTSPS covers critical obstacles to OLPC deployment in an optimal way.

Some detailed info may be available through RUPP and/or IRAP (via DoLIDAR). Perhaps only in Nepali.

 

Tools

Project Management

Consider TaskJuggler

.

Distribution

Need reliable (tested, failsafe) automatic remote updates from release channels via email/news - both for servers and CardPodCaster firmware. Consider use of mrepo

for repository.

 

Cloud Computing

Scalable central servers. Play around with various options testing with virtualization, LVM and cloud computing. Appliance Tools

in Fedora 10 may be helpful.

Consider AMPQ subset of MRG enterprise message queues

now in Fedora 10.

 

Literacy and Broadcasting 

 

 Nomadic women in Mongolia face extreme problems for literacy programs. Print, plus broadcasting plus visiting teachers proved successful in the Gobi women's project. No mention of cassette recorders, perhaps because even radio required distribution of 100 batteries per receiver.

 

CardPodCasters

A CardPodCaster is a cheap MP3 player that can play from and record to an SDHC card instead of using internal flash. It still has a USB connector so can also be used as general purpose SDHC card reader.

 

So far have only been able to find CardPodPlayers (without internal mike and record capability).

 

Oddly the one I have (Olin OM106) does not support recording even though it includes FM radio.

 

Also the FM isn't very useful. Needs better earphone antenna, has annoying level of hiss and won't record presets as claimed in manual.

 

This may indicate some problem with other easily available existing models using same chipset. May need hard work, ECO or ODM to source model with line input and/or internal mike.

 

LCD screen only useful for setting FM channel (multiple clicks) and showing track number and total. Elapsed time within track also shown but useless since cannot position within track but only to next and previous track. No filename or id3 tags shown. No folder navigation.

 

Have not yet checked whether only uses root directory or also sequences through full tree. If so whether breadth first or depth first and whether alphabetical order of file and folder names (unlikely). Anyway the screen only provides number within the sequence so you still have to just press forward back the appropriate number of times to get from one track to another whether it accepts folders or not. Slightly cheaper model with no screen at all worth considering (but if recording only comes with FM capability for very little extra then LCD screen may be desirable).

 

However confirmed that does work fine as an SDHC card reader. Am typing this from Fedora 9 installed with no problems on laptop with no HDD using 8GB (class 6) card in card reader plugged into USB.

 

LCD blue backlight always on when plugged into USB (displays "USb" and disables all buttons). Similar backlight always on issue for car model plugged in to 12V cigarette lighter socket. Reasonable behaviour in both cases since can assume plenty of power available from USB supply or car battery. But will probably want firmware change to conserve battery if just using external USB or 12V portable battery pack. (Also firmware and/or ECO to permit accepting power from USB port without switching to USB card reader mode).

 

SD Cards

 

Should check all types of FAT format SD and SDHC cards for compatability. Only cheapest (slowest) needed, not faster class 6 as recommended for computer use.

 https://www.redhat.com/archives/fedora-olpc-list/2008-November/msg00041.html

On-disk.com supplied card duplication and labelling and order fulfilment for Fedora 10 on OLPC as well as general CDROM and DVD handling business for FOSS developers. Karlie Robinson might be source of advice on issues like supplies, labelling, hassles etc.

 

External VGA

Absence of VGA port is an important limitation for XO use with projectors and larger LCD displays as would otherwise be feasible in areas with sufficient electricity.

This is mitigated by using a USB to VGA adaptor (at extra cost). LCD displays (and presumably projectors) with this built in are also becoming available.

Requires models supporting chipsets eg "bluebox" brand and software configuration as recommended here.

 

 Postal

 This project for distance education with DVDs by post in India looks worth careful review.

Digital StudyHall (DSH) is a research project that seeks to improve education for the poor children in slum and rural schools in India. In a nutshell, think of its technical approach as the educational equivalent of Netflix + YouTube + Kazaa.

We digitally record live classes by the best grassroots teachers, transmit them on the "Postmanet" (effected by DVDs sent in the postal system), collect them in a large distributed database, and distribute them on DVDs to poor rural and slum schools. Education experts and teachers use the system to explore pedagogical approaches involving local teachers actively "mediating" the video lessons. By harvesting a "viral phenomenon" of community participation, DSH aims to help train teachers and deliver quality instruction to underprivileged children. The project is a collaboration between computer scientists and education experts. The main aspects of DSH are:

 

A live deployment of DSH has been operating in India since the summer of 2005. As of summer of 2008, we run pilot "hubs" in Lucknow, Calcutta, Pune, and Dhaka, covering approximately 30 schools. And during this time, we have accumulated more than 1000 recordings of lessons in English, math, and science, in Hindi, Bengali, Kannada, Marathi, Tamil, and English. We have also started applying the same approach to agriculture extension work (Digital Green) and awareness campaign for rural healthcare (Digital Polyclinic, operating in Lucknow and Ghana). Today, DSH is still a young research project, as we continue to work on rigorous evaluations and seek to understand many outstanding questions. We have, however, already seen initial promising signs, and we hope to eventually scale up the system to cover a far greater number of children, contributing toward the Millennium Development Goal of universal primary education.

Digital Study Hall

 

 

 

 

 

 

 

 

 

Document Actions
Log in


Forgot your password?
New user?