Solution: Use a HP PCL driver instead. This will result in 3-4 s print times for the aforementioned DIN A5 label. This problem should affect all operating systems, as one can configure all of them to use a PostScript driver for printing to this printer. However, it will rarely affect Windows, as most users will use the manufacturer supplied driver there.

To use a PCL driver in Linux, choose the hpijs driver (named “HP LaserJet 2200 Series hpijs, 3.11.7″ in the driver database). Be sure to change “Printer Options: Printout Mode: Normal” to “High Quality” in the printer default properties, accessible from the printer management system tool. Otherwise, the rendering quality will be visibly bad. Even with the “High Quality” setting, this driver will only print with 600 dpi, while the Postscript driver will print with 1200. But that difference is hardly ever noticeable without a magnifying glass.

An alternative is to use the gutenprint-ijs driver (named “Generic PCL 6 PCL XL LF Printer Foomatic/gutenprint-ijs.5.2″ in the driver database). Likewise, set it to “Printer Options: Printout Mode: High Quality” in the printer default properties or it will even have a worse, “fringy” rendering quality than the hpijs driver. It also gives 3-4 s print times for the DHL Intraship labels, but also only achieves 600 dpi.

To use a PCL driver on Mac OS X, use the “Generic PCL” printer driver, not the “HP Laserjet 2200″ one. [TODO: It has to be confirmed that this helps.]

Discussion:  This printer understands PostScript Level 2, but only as an emulation. This probably makes processing slow for some elements, maybe the barcodes on shipping labels. Talking to the printer in its native PCL6 language fixes this.

Changes that did not help:

• This problem could not be fixed by choosing “print as image” in the Adobe Reader options. This works, but is even slower than the 125 s print time for a DHL shipping label.
• This problem could not be fixed by sending  Postscript Level 3 instead (the LaserJet 2200 only understands level 2, it will not even blink for transmission, and the print job is discarded after a time).
• This problem could also not be fixed by using a Postscript Level 2 driver and adapting the settings from “Resolution: FastRes 1200″ to “600 dpi” and from “Resolution Enhancement: Printer’s Current Setting” to “Off”. Print time was exactly the same 125 s for the DHL shipping label, and print quality was even with magnification hardly any different from the original printouts. Maybe only the “Resolution Enhancement” was disabled, while the “Resolution” setting does not work at all?
• This problem could not be solved by setting the “Total Printer Memory” parameter from “8-15 MB” to “16-23 MB”. (The printer had 16 MB and the idea was, maybe only 8 MB get used with that current setting, as the computer probably cannot determine how much is in the printer.) However, this did not improve printing speed by any means. Print time was exactly the same 125 s for the DHL shipping label.

Changes that might help but have not been tried:

• Pre-processing the PDF label document with a command like ps2ps. This might convert the Postscript elements to others that take less effort to render.
• Adding more memory to the printer. Because maybe the “difficult” Postscript elements just need more memory, not more processing power. And when lacking the proper amount of physical memory, this printer will will automatically compress memory content to be able to process the document [source: HP Manual for this printer]. And this processing will make it take more time, probably.

Currently, there seems to be no way to enable the full 1200 dpi using a PCL driver. The problem is also reported on the driver’s support page.

First name is the real table name in M2E Pro 2.0.13, second name is a more intuitive one proposed by me.

• m2e (m2e_tmpl_projects): Contains listing templates, corresponding to the backend concept of the same name.
• m2epricetemplates (m2e_tmpl_prices): Contains price templates, just as they can be modified in the backend. Note that listing templates contain the price template information redundantly.
• m2etemplates (m2e_tmpl_descriptions): Contains description templates, corresponding to the backend concept of the same name. Note that listing templates contain the price template information redundantly.
• m2e_account (m2e_licences): Contains the licences for the M2E Pro software.
• m2e_account_list (m2e_ebayaccounts): Contains eBay accounts used in M2E Pro.
• m2e_api_versions ():
• m2e_calculated_shipping ():
• m2e_category_version ():
• m2e_currency (): Contains currency abbreviations and their long form.
• m2e_ebay_categories_77 (): The eBay category tree with category numbers, as fetched from eBay.
• m2e_items ():
• m2e_item_specifics ():
• m2e_listings (m2e_projects): Contains listing sets: the things visible immediately when clicking on “M2E Pro -> Listings”. They are called “projects” in the database structure.
• m2e_listing_item_shipping (m2e_listingtmpl_shipping): Contains the shipping methods of listing templates.
• m2e_log (m2e_log_projects): The log of projects, containing eBay API results from listing, deleting etc. the articles of projects.
• m2e_logcleaning_settings (m2e_config_log): Settings when logs should be cleaned automatically.
• m2e_marketplace (m2e_marketplaces): Definitions of international eBay sites.
• m2e_product_to_project (m2e_listings): The actual products contained in projects (“listing sets”), together with their linked eBay items if applicable.
• m2e_p_to_p (m2e_listingtmpl_payments): Contains the payment methods of listing templates. The current name suggests to mean “payment to project”, but that would be wrong: projects are the actual listing sets, not their templates.
• m2e_shedule_task_settings (m2e_config_sync): Configuration saying which sync tasks are enabled.
• m2e_store_management (m2e_config_syncwindow): Configuration which start time to use for the time window when fetching updates from eBay. It will be 24 hours before the last successful sync.
• m2e_synchronization_log (m2e_log_sync): The log of syncing with eBay, as available in “M2E Pro -> Logs -> Synchronization Logs”.
• m2e_templates_descriptions (m2e_tmpl_ebay): eBay’s article templates, as downloaded from them.
• m2e_transactions_events (m2e_log_transactions): The transactions done via eBay, as available in “M2E Pro -> Logs -> Transaction Log”.

Basically, there are three groups of vee-belt and other belt profiles in use in Europe today:

• conventional or classic vee-belts (German “klassische Keilriemen”): standardized in DIN 2215 / ISO 4184; using one-letter profile names in different sizes (Z, A, B, C, D, E), in many cases the profile is also named by its width in millimeters (10, 13, 17, 22, 32, 40).
• narrow-profile vee-belts (German “Schmalkeilriemen”): standardized in DIN 7753 Part 1 / ISO 4184; using SP profile names in different sizes (SPZ, SPA, SPB, SPC); maybe, “SP” means German “Schmalprofil”?
• high-performance narrow-profile vee-belts, open shoulder, toothed (German: “Hochleistungs-Schmalkeilriemen – flankenoffen, formgezahnt”): standardized in Europe in DIN 7753 Part 1; using XP… profile names in different sizes (XPZ, XPA, XPB, XPC)
• wide-profile vee-belts (German “Breitkeilriemen”, “Variatorriemen”): standardized in DIN 7719 / ISO 1604
• flat belts (German “Flachriemen”): used in different applications such as tangential belts, folding and conveyor belts and machine belts [source]. There seems to be no widely accepted standardization in this area.
• round belts (German “Rundriemen): mostly used for conveying and driving tasks in mechanical engineering.

For a short description of the different types, see HUG-Technik on Keilriemen [German]; for a more detailed introduction, read tedata.com on vee-belts. For an overview about DIN and ISO standards relevant for vee-belts and similar devices, see HUG-Technik on important standards for belts [German]. See also the English Wikipedia on mechanical belts and the German Wikipedia on vee-belts; though both of them lack extensive information on belt specs yet.

The DIN and ISO standards cited above are used in Europe; in the US, the standard RMA/MPTA (and in the UK, BS 3790) specifies minimally different but overall compatible profiles with other profile names. [source]

Standards for vee-belt length

Length conversion table. A very important tool to work with belt specs is a vee-belt conversion table. That’s because the nominal length is different for the different types of belts (for example, it is “inner length” for classic vee-belts – also the only error in the linked conversion table).

Li, Lw, La. For every vee-belt, one can give three lengths: inner length Li, effective length Lw and outer length La. Inner and outer length are the inner resp. outer circumference measure of the vee-belth, without any linear tension on the belt and in circular shape. Effective length is a fictive median length of a vee-belt that is the circumference at a certain depth of the belt profile; cf. column “effective width” (German: Richtbreite) in the vee-belt conversion table. Lw is used as the nominal length of belts with profiles SP*, XP* and X*. Synonymous to Lw, some manufacturers use Ld or Lp. [source]

Belt number. For classical vee-belts, there is another identification system in addition to the “20 x 3500 Li” type: the belt number (German: Riemen-Nr.). It consists of the normal profile size designation letter and a number that normally corresponds to the inner length in inches (rounded, where necessary); for example, “Z 22″ or “D 150″. [source] A list of these numbers can be found on this vee-belt index.

Measuring vee-belt profiles

All values below are measured with light touch of the caliper.

• height of V-belts
• nominal 10 mm; measured 9.3 – 9.7 mm (in convex curves, 9.8 mm)
• width of V-belts
• nominal 12.5 mm; measured 12.3 – 12.4 mm

Measuring vee-belt length

Measuring vee-belt length Lw. It is usually proposed to measure a vee-belt by cutting it and nailing it flat to a board [source]. However, this is not what you want if you need to determine the size of a new vee-belt that has lost its labeling. So here is a different procedure that I tested successfully:

1. Mark a line on your flat, hard floor by taping 2 measuring sticks to it and also glueing a stop to the start.
2. Add a small cable tie around the vee-belt to be measured and use that for marking the start and aligning it to the start marker you glued to the ground.
3. Roll your vee-belt on the ground along the measuring sticks until you went one full circle.
4. Take the measure and interpret it as effective length Lw of the vee-belt (which it is, approximately).

You could do two or three measurements and take the average, but this seems not necessary as this kind of measuring has repeatable results that are up to 1 mm exact. Measuring with a flexible measuring tape while holding the vee-belt in your hands is not recommended, the results are less exact (my experience: measurements of a single belt resulted in 3505 mm Lw and 3520 mm Lw, while the result with the roll-on-floor technique was 3522 mm Lw). When measuring while holding the belt in your hand you have to take special care to not bend it while measuring, as bending will increase the measurement more towards La; in this example, from 3522 mm Lw to 3533 mm for lightly bending).

It is said that the measure taken by this procedure is the median length of the vee-belt [source], which probably should mean, the average of inner and outer length. In practice, it is said to be a good-enough approximation of the effective length Lw. While this is still true, my tests suggest that the length measured this way is not always the median length, but “something between inner and outer length”; exactly which mostly depends on how the belt is built, namely, how far to the outside the pull-resistant strings are located, as these seem to work like a hinge when bending or unbending the belt, affecting the depth of the area on each side that gets either compressed (so, shortened) or pulled on (so, lengthened). These fibers are normally located right below the upper edge of the belt, so the measure is normally more towards the outer length than the Lw measure is.

Measusing vee-belt length La. It is said that one can do that by placing the belt in circular shape on a flat surface and placing a flexible measuring tape around it [source]. However in practice, this is hard to do as you need some tension on the tape to not measure too much. It works well when the velt is still mounted, though.

Determining Lw without a belt. In case you have a device needing a belt but don’t know which one, place a rope around the belt path that has approx. the thickness of the pulley nuts. Mark the length of rope you need, take it out again and measure that length when laying straight on flat ground. This is a  good approximate for the effective length Lw of a fitting belt. (As an alternative, there is a forumla to calculate this, and online calculators for that.)

Various experiences with measuring vee-belts

• At times, some vee-belts seem to use the wrong signing schema. One belt had the classic profile, so the sigining of “20 x 3550″ was to be interpreted as Li = 3550 mm. However, all measurements turned out with Lw = 3522 mm, so an even larger measure of 3550 mm can only be La, not Li. The formula produces a result coherent with this: La = Lw + 31 mm = 3522 mm + 31 mm = 3553 mm. Another, quite old Continental belt was SPA profile so should be labeled with Lw, yet the label said “12,5 x 1200 La”.
• Amount of difference between measured flat length and Lw. In one case, a classic 20 x 2000 mm Li belt was measured with 2060 mm flat length. Lw according to the formula is Lw = Li + 48 mm = 2048 mm Lw. The difference of 12 mm isquite small and probably comes from the fact that the fibers are not exactly located at the diameter corresponding to Lw.
• Different formulae for converting to and from Lw. Interestingly, there seems to be at times a slight divergence of what formula is to be used for converting from effective length Lw to Li and La. For example, this vee-belt conversion table and basically all other such conversion tables on the web state for a 17 mm wide classical vee-belt: Lw = Li + 40. However, one 17 mm clasical vee-belt belt was found with an inscription saying “PETER-BTR 17 x 1320 Li / 1363 Lw”, corresponding to Lw = Li + 43.

Storing vee-belts correctly

It is said that vee-belts are normally built by manufacturers to reach a lab runtime of 25 000 hours [source]. If this is reached in practice also depends on proper storage conditions.

If properly stored, vee-belt properties do not change for several years. However, most rubber-based products will deteriorate if improperly stored or handled (like being exposed to oxygen, ozone, extreme temperatures, light, humidity or various solvent agents). [source]

Therefore, the storage location should be dry and dust-free and must not contain chemicals, oils or solvents at the same time. Vee-belts should be stored without any force on them (pressure or pulling force) to avoid any permanent shape changes and other damaged. This also means that they should only be stored in hanging condition if the hanger pin is at least ten times the height of the belt profile. Additional maintenance hints for rubber products are found in DIN 7716. [source]

This made me get on an idea about the (yet largely unsolved) meaning and role of dreams: namely, to develop and to train the raw material and the speed for flow state thinking. Because dreams go much faster than the real world, and flow state imaginary thinking of “what probably happens next” is just like that, much faster than reality and than normal thinking.

You know “flow state” from these horrific little moments when you see something bad happening to you physically some 0.5 to 2 s before it happens indeed. These seconds feel like slow motion: you think so much in this time about what happens and what to do to avoid it that you later wonder how it all did fit into these fractions of a second. That’s because thinking is much faster in this state of mind – and because reality provides no training to react to such situations, maybe the fast pace of dreaming is this training for the dangerous situations in life, training our mental abilities to react fast enough to avoid the worst outcome.

All speculation of course, but perhaps somebody feels inspired to do some research

Now, these both ways to live life could not stop the deterioration of Western societies, which started in (say) the mid 1990′s and is in good full swing as of 2011. It’s not the technichal challenges of more scarce resources that we face, or the dangers of unregulated greed (in the financial markets and at home). The real, real problem is that in a land of Cockaigne like this, parents can never succeed to transfer their builder mindset (of postponed consumption, investment, hope, betterment) to their children and grandchildren. These will find out that near everything can be had without work, and will get too comfortable to even be interested, to desire to learn, and to be creative. It’s bad with the children and worse with the grandchildren. Call it decadence.

Now what? Let’s doubt that taking part in decadence is never a life lived for what’s adequate, whether you have children more decadent than you, or no children, being decadent yourself. Instead, there has to be a life to be lived against decadence. Successfully. What do we need for that?

The new, new frontier. The US could only start to become decadent where the frontier for settlement had passed, leaving civilization, but without vision and without tasks. Of course, people found something to improve and to do during some of the following generations, but finally even that faded. There was something like a last try to remedy this by calling the race for the moon the “new frontier”. Nice try, but itself influenced by decadence: Is space travelling really the first thing that comes to mind when you think of “task” and “vision”?

What we need is the new, new frontiere. And then the new, new, new one. And so on. We need eternal pioneering. Look around and see that the world is in nearly no place as it should be. Sahara is a desert, the Congo struck by malaria, all of Africa by Aids, and much of the world by wrong belief systems. Just for the start. And we Westerners dare to say we see no tasks at hand, when questioning ourselves how to recreate vision and zeal in our society, or when questioning ourselves what to do with the >20% jobless in southern Europe? And then people dare to get on the idea that adding another child or two or three to already 7 000 000 000 people on Earth is a proper reason to live for, even though its more-than-apparent that even more people will even intensify the problems and resource conflicts on this planet?

Let’s get practical: we know how to do military operations, with respect to human resources. People get “deployed” in groups of several hundreds to some battlefield overseas, for some 6-8 months maybe. Then they are exchanged by a new group, and themselves go back to their home country to refresh and prepare again for their next deployment in another 6-8 months. Let’s take this over to all kinds of development aid. Because this is attributing the right role to civilization and homes: a place of refreshment and preparation for your “mission”.

Specifically, I propose to create settlements of ca. 4000 people, who travel to overseas development assistance deployments in groups of 1000. As this is no military operation, this can and should include the whole families (children can get schooled underway). The people at home would work to support them, from logistics to medical care etc.. As in military deployments, there will have to be like 4-5 (minimum) immediate support people behind everybody on a mission, and even more so working for money for these missions. But in any case, everybody’s goal should be to go on deployment (and if only for 6 in 36 months), being the reward for year-long support work.

Some ideas for specific missions:

• Re-afforest the whole of Sahara. Yea, we had that above. Just plant the trees, it’s good for the planet
• Build some big islands, or enlarge others. Having more land in nice climate is always a good thing. Why, for heaven’s sake, would people want to live in Siberia while there’s an island waiting for them?
• Eradicate AIDS.
• Eradicate malaria.
• Eradicate hemorrhagic fevers. And all those other infection sicknesses.
• We want the rain forest back where it was!
• We want whales back, and the Indian elephant in big numbers, and all these other lovely creatures. Isn’t it somewhat … heavily self-conceited, that we allow 7 billion humans on Earth and reduced many of the other  creatures to just some thousands per species?
• Yea and finally, once we repaired and tidied up the Earth, there are other planets to travel to. Yet, first things first!

Remember what the Lord God said after the deluge: spread and multiply and fill the Earth. But Josephus says, the people wanted to stay together in one location, building Babylon and its towever as their central collection point. Yet God wanted them to go and colonize the Earth, he wanted them to pioneer – maybe because of the danger of rapid decadence, should they stay together in all their civilization? So he had to mix up their language …

Pioneering never ends. It must not end, as it’s a crucial ingredient for human mental health.

It’s simple:

1. In the command line, use the adduser command:
   adduser somename
2. Do NOT enter that user’s name to /etc/ftpusers, as this file lists users NOT allowed to log in by FTP.
3. There is no need to restart proftpd, as it will dynamically react to changes in your set of system users. If you still need to restart it (to read other config changes), you need to send it the SIGHUP signal [source], as it’s managed by inetd (the Internet super daemon, a kind off automatic start / stop service for other daemons, so that there is no need and no option to restart proFTPD manually).

The official docs are here: proFTPD Mini HowTo Index. Please esp. refer to the ProFTPD Logins and Authentication in case your new user still has problems logging in.

The main problem, is are these SpamAssassin filtering rules, as indictaed in the e-mail headers:

1.1 FH_HELO_EQ_D_D_D_D Helo is d-d-d-d
1.3 RDNS_NONE Delivered to internal network by a host with no rDNS
3.2 HELO_DYNAMIC_IPADDR Relay HELO'd using suspicious hostname (IP addr 1)

The HELO_DYNAMIC_IPADDR rule

This is the main problem in this case, contributing 3.2 points. The SpamAssassin rule descriptions for this rule do not exist on the rule list page. But judging from similar rules, it appears to mean this:

“An untrusted relay uses a hostname in HELO that indicates a dynamically allocated IP address. (So this is different from RDNS_DYNAMIC by referring to the name used in HELO, not the RDNS record, though noth might be the same name.) Legitimate mail servers would usually be expected to have static allocations. This is determined by checking the reverse DNS name of that relay against a number of inofficial naming conventions. Reverse mapped names that contain sections matching strings such as “dynamic”, “dhcp”, “adsl” etc. may be matched by this rule, as may names that contain strings of numbers (like lvps165-624-87-23.example.com).”

The following sources have been used: SpamAssassin RDNS_DYNAMIC rule page and HELO_DYNAMIC_DHCP rule page, a bug report and a forum thread. With this interpretation it became clear that the SpamAssassin rule was triggered by the generic hostname used in the virtual server’s HELO, as indicated by this mail header (IP addresses are randomized):

from [43.151.120.12] (port=50581 helo=lvps43-151-120-12.dedicated.hosteurope.de)

To fix this, follow these steps:

1. Think of a full qualified domain name to use as your hostname; for example, server.example.com, but not just example.com. The RDNS name you will set up later may or may not be the same as the server’s hostname; for simplicity, I recommend it is the same.
2. Configure your DNS zone to resolve that FQDN to your server’s IP address. You will probably do this with Plesk on the Host Europe server (in Plesk 9: Home -> Domains -> [select domain] -> DNS settings). But maybe you use an external DNS service.
3. Follow the Host Europe instructions to change the virtual server hostname permanently.
4. Ensure that this new hostname is used for HELO when sending mails. For that, just check how your SMTP server application on the virtual server identifies itself when contacting it with telnet on port 25 (for example: “telnet 74.161.120.13 25″) [source].

 The RDNS_NONE rule

This rule is triggered because a vanilla Host Europe virtual server does not yet have RDNS resolution. (It also is still triggered when configuring a RDNS name, as for proper RDNS resolution that name itself has to point back to the IP address used as input for the RDNS query.)

Host Europe knows about this, as they say in their FAQ that some e-mail servers will even refuse to accept e-mails from servers that do not have proper RDNS resolution [source]. And there are reports by customers at other hosting companies, reporting that incorrect RDNS setup indeed contributed to problems with mails arriving as spam [example].

The requirement for proper RDNS resolution is that an IP address resolves to a PTR (“RDNS name”) that itself resolves to that same IP address (via an A record). How to set this up on a Host Europe virtual server:

1. Choose a fully qualified domain name as your RDNS name. We proposed above to re-use your hostname as the RDNS name.
2. If you did not chose it to be the same as your hostname, make it resolve to the server’s IP address. In the case of your hostname, you have already done that above.
3. Follow the Host Europe instructions to change the RDNS name in the customer information system.
4. Wait for your changes to become effective. This may take 48 hours or a bit longer to propagate through the whole Internet [source].
5. Check that your RDNS resolution works correctly now. According to [serversupportforum.de thread 124291 post 11 and thread 124324 post 12], do this:
1. Check resolving from IP address to PTR (“RDNS name”): dig -x 217.79.215.140
2. Check resolving from the determined PTR (here, bundestag.de) back to IP address (here, must be 217.79.215.140 again): dig bundestag.de

• Wires cannot be annotated with name/value pairs in KiCAD. But you can create a point-type “wire meta component” yourself, maybe looking like a small dotted circle line. You can then place it as the center of every logical wire, and connect its both pins to the pins of connected devices. Now you can add meta information to your wire, and also display these as fields in the logical middle of your wire: wire color, crosssection, cable type, special requirements for chemical and heat resistance etc..
• KiCAD will display all wires in the same color. So in case you rely on wire color for wire identification in your project, the resulting wiring diagram will be hard to read for (visually) for working with your physical project, even though the wire color can be displayed in a field (see previous tip). For that reason, better use wires only to encode the general role of a wire (like red for all positive power supply wires, black for the negative ones). And rely on wire ID numbering for identification purposes. There are many commercial products for adding these numbers to your wires also after installing them (like cable flags, or as a more durable solution, cable ties with 10×2.5 mm label field). You can add these wire identifiers to the fields with the pin descriptions of your components, maybe using a vertical bar “|” to separate wire ID and pin name / number, and place the wire ID on that side of the field more to the edge of the component.
• Even if KiCAD could display different wire colors, wire IDs would still be needed if you use buses, because a bus could contain two wires of the same color so that tracing them in the diagram would be impossible. And you should use buses, they are a great means for tidying your wiring diagram up optically. Especially, use buses for all multi-core cables in your application, as these are a sort of “physical buses” anyway.
• Use the KiCAD “Edit -> Find” functionality to find dsired elements in your wiring diagram quickly; this adds a good deal of interaction to your diagram. You should design your IDs for wires and components, and everything else that appears on labels in your physical product, in a way so that every such string is unique. This makes searching through the diagram more comfortable, as you know that the first (and only) result is the one you are looking for.
• Look through kicadlib.org to see if you can locate useful components not already in your locally installed KiCAD libary.

My recommended fonts

• DejaVu Sans. Its web home is dejavu-fonts.org. Recommended to be used as the main font for the document (except you insist on a seirf font). Because, this font has great Unicode coverage and free software licencing, and as a unique feature among comparable fonts like FreeSans, it comes with  a “Light” variant. What is really missing for now is a “Light Condended” variant, however.
• Liberation Sans Narrow. Even more condensed than Liberation Sans, which is already more condensed than DejaVu Sans but comparable to DejaVu Sans Condensed. So it can be used as a “really condensed” font in replacement of a missing “DejaVu Sans Light Condended”, increasing the possibilities of document formatting. The downside is, this font lacks good Unicode coverage, so it can only be used for normal Latin text, not formulae etc..
• FreeSerif. A good-looking, free software licenced serif font with huge Unicode coverage.

My not-so-recommended fonts

In alphanumeric order.

•  CM-Super Sans. Like Modern Sans, this is also a PostScript Type1 conversion from the TeX Computer Modern font. And therefore, it is really good-looking. However, if any of these TeX fonts would be used, Latin Modern Sans would be preferrable.
  Because its Ubuntu package lmodern says in the description: “Their size is reasonable and they are usually considered to be of good quality (compared to cm-super, for instance; however, cm-super contains font families that have no equivalent in this package; additionally, there are character sets that are supported by cm-super and not by the Latin Modern fonts).” In addition, Wikipedia confirms this: “Other PostScript-based replacements exist such as BaKoMa, CM-super, or Latin Modern, instead of the original METAFONT-based Computer Modern. The Latin Modern implementation, maintained by Bogusław Jackowski and Janusz M. Nowacki, is now standard in the TeX community and was made through a METAFONT/MetaPost derivative called METATYPE1.” [source]. An advantage over the Latin Modern fonts is however that they contain fonts for all font sizes [source].
  Corresponding Ubuntu packages: cm-super, cm-super-x11. But for some reason, this does not make the sans serif font available to X11, or maybe not even install it.
• DejaVu Serif. For my taste, this font has a very bad look for a serif font.
• FreeSans. Comparable in good look to DejaVu Sans; FreeSans is just a bit more condensed. Comparable Unicode coverage as DejaVu Sans. However, this font lacks a “Light” variant.
• Latin Modern Sans. This is a PostScript Type1 conversion of the LaTeX Computer Modern font. As it’s coming from TeX, it’s definitely the best-looking free software licenced sans serif font [text examples]. However, it is not recommended here because it lacks sufficient Unicode coverage.
• Corresponding font names in LibreOffice: LMSans, LMSansDemiCond, LMSansQuot.
• Corresponding Ubuntu package: lmodern.
• Liberation Sans. This font is a bit more condensed than DejaVu Sans. Compared to DejaVu Sans, it misses the great Unicode coverage and also a “Light” typeface, which is the reason it was not recommended here. It has larger relative x-height than DejaVu Sans, which is said to make it less useful for large portions of text [source].
• Liberation Serif. Good-looking, screen readable serif font. Already using another font from this family is one reason more to use this serif font instead of adding a font from a third family. However, this font lacks big enough Unicode coverage, and cannot be used for that reason.
• Linux Libertine G. It is a way better looking serif font than DejaVu Serif, includes considerable Unicode support, but is less readable than Liberation Serif in small font sizes.
• Lucida Sans. This is not recommended here because it is not available under free licences; it is just included gratis with Java. What’s interesting however is the Lucida Math font for formulae.
• Nimbus Sans. Not recommedned to be used as teh default font because of lacking Unicode support and lack of a “Light” font weight in the free software licenced variant [source].

Technical note: If you are on Linux, do not check Unicode coverage of fonts with gucharmap, as it will simply take over non-available symbols from fonts that have them. Use the LibreOffice “special character” table instead.

Technique using ntfsclone

ntfsclone has the advantage that it can supply zeroes for unused sectors on the fly, eliminating one step of the dd based process (see below).

1. Start from a Linux live CD, such as Ubuntu, and execute this command (naming your hard disk to backup, here /dev/sda is used):

   sudo ntfsclone -o - /dev/sda | bzip2 -c > Backup.yyyy-mm-dd.sda.raw.bz2

2. Note that you can also back up just individual partitions. But then, you should back up the MBR and boot loader separately, and the partition table also in text format. For that, see:
   • Transfer Windows to New Hard Disk with ntfsclone
   • Notes on backing up entire hard disks or partitions
3. For uncompressing later, use a command like this:

   bunzip2 -c Backup.yyyy-mm-dd.sda.raw.bz2 | ntfsclone --overwrite /dev/sda -

Technique using dd and gzip

This involves an additional step for manually zero-ing unused disk space. Also, using gzip instead of bzip2 yields less effective compression, but is faster. You can however supply bzip2, bzcat and bunzip2 for every call to gzip, zcat or gunzip; their parameter usage is compatible for our purposes.

1. To achieve good compression ratios, it is advisable to fill unused space of the disk to back up with zeroes (or ones or any other kind of redundancy). Please do not do that when trying to rescue defective disks and partitions, as you would lose potentially valuable data in unallocated space on your original disk. There are several techniques for that:
• In any kind of file system, when using Linux, you could create a file that is as large as possible, consisting only of zeroes, and after that delete it again:
  dd if=/dev/zero of=fillfile.raw bs=8M; rm fillfile.raw;
• If it’s about a NTFS partition and you can use that in Windows, you could use Microsoft’s SDelete. However, when trying it out in 2008 by using “sdelete -c -p 1 c”, it gave me this error message: “Cannot clean free space for UNC drive.”.
• Again, if it’s about an NTFS partition in Windows, you can use Eraser. In version 5, you could configure an own pattern (just zeroes) to overwrite unused disk space. In version 6.0.8 at least, this is no longer immediately possible. The developer advises to use the “HMG IS5 Baseline” for that, as it would do just that task of overwriting once with zeroes [source]; but in my test, this failed (no meaningful compression ratio afterwards). You can still create an own erasure method according to page 15 of the user manual though [source]. Also note, to be able to erase unused disk space when running it in Windows Vista, you need to run Eraser as Administrator user. For that, close all Eraser instances (no Eraser tray icon must be left!), right-click the Eraser desktop symbol and click “Run as Administrator …” [source].
2. Now connect the hard disk to back up as a second or external USB hard disk to a Linux system, boot its installed Linux system, or boot into a Linux Live CD such as aUbuntu.
3. To back up the hard disk to a gzip-compressed image, use this command (naming your hard disk, here /dev/sda):

   dd if=/dev/sda | gzip -c > Backup.yyyy-mm-dd.sda.dd.gz

   On a 1.3 GHz Pentium 4 processor, this produced a throughput of between 3 MiB/s of hard disk raw data (at the start of the hard disk where compression was difficult) and 10 MiB/s (at the end of the hard disk where the all-zero area was located).

4. Using “kill -USR1 <pid>”, supplying the process ID of the dd process, you can look at the progress of the command.

5. For restoring from the compressed backup, you would do this:

   sudo zcat Backup.yyyy-mm-dd.sda.dd.gz | dd of=/dev/sda

   Or if you want a more exact pendant to the compress command, use this:

   sudo dd if=Backup.yyyy-mm-dd.sda.dd.gz | zcat | dd of=/dev/sda

6. Using the “seek” option of dd, or using cat, it should be possible to split the created file into several ones (for example, to burn them to several DVDs). However, I did not try that out. What definitely works is burning one backup file to a DVD-ROM (e.g. a 8 GB double layer DVD-R). Previously, there were some fixable problems with files >4 GiB on DVD.

Non-working alternatives

The original problem had been rather to back up a complete hard disk using Windows on a  ThinkPad T40. There were two options availale using IBM software, but both did not work (so the above method was used instead):

• The software “ThinkVantage Rescue and Recovery” should normally work as a solution. However, version 4.1 of it refused to install on a ThinkPad T40 because it thought the operating system is Windows Vista rather than XP.
• Alternatively, “IBM Rapid Restore Ultra” from “Access IBM” should be a possible solution. Unfortunately, this is not possible either, because: to create the backup partition, a floppy drive is needed and it was not possible (even not by reinstalling that application) to select the option that would make a USB drive sufficient for creating the backup.

A financial-economic crisis (like this or this or this) makes us think what’s wrong with our economic systems, and search for alternatives. For example, one would search for grassroots ways to organize mutual service and exchange of goods without using legal tender, as the latter can be a scarce resource due to “personal deflation”, esp. in an economic crisis (that is, by being separated from any high-capacity money source of the society, e.g. by being self-employed and having only “poor” customers).

Interestingly, many alternative economic concepts have been developed long ago at concept level, but failed to receive public attention or achieve widespread application. For example, in the area of trade mechanisms and finance, I found several concepts that run for being a basis of an alternative economy:

• LETS, the Local Exchange Trading Systems. Effectively, these are complementary currency systems on a local scale, where members exchange goods and services for LETS credit points. All members start with zero balance, and go below that by paying with credit points; which means, they have the role of a bank that does bank money creation. The value of goods and services is something to be agreed by the members, that is, normal market mechanisms of supply and demand apply. In principle, all LETS systems are confined to one city or small area (see e.g. this LETS list from 2002), as the credits are backed by just group trust (omnidirectional trust from one to all other LETS group members). This is, in my view, their biggest shortcoming. How do you transfer your LETS credits when you move, and even worse, how do you use LETS if you have a nomadic lifestyle? Also, the island-type of LETS, with no standardized infrastructure between them, severely limits its public visibility as one movement or system.
• Time-based currency (see also time banking). Also called “time dollars”. While the bookkeeping is very similar to LETS, the difference is that noarbitrary unit (or a unit aligned to legal tender) is used to measure service contributions, but the man-hour of work time. This can, but needs not, be used to build egalitarian systems, where everybody’s contributions are valued equally. For an interesting early experiment from 1827, read about the Cincinnati Time Store. Currently, timebanks are normally local organizations just like LETS [see an international list of time banks].
• Barter platforms. By these, I mean systems that are based on direct barter between two parties, or circular barter between three or more parties. But always so that a deal happens only when all parties at the same time have something that they can offer that another party in this deal wants. That is, they work without intermediate units of accounting, which would eliminate the need that (1) supply and demand are matched for every party at the same time and (2) somebody can only be part of a deal by both demanding and supplying something, as none of these can be replaced with currency. These severe limitations of direct barter make these platforms look quite unusable, esp. before reaching a critical mass of offers. Examples of such systems include the German portal Bambali.
• Scrip. Alternative currency by simply writing some numbers on some paper.
• Mutualism. An economic theory that seeks to value goods just by the amount of labour / resources needed to produce them, thus eliminating the concept of capitalist “gain” (price above cost) altogether.
• And there is a list with many more articles on alternative monetary systems.

Interestingly, the proposal for a time-based currency even made it into a United Nations document (but note that the document was compiled by members of NGO’s when meeting for the Millenium Forum, so is no official UN resolution or anything):

“The Forum urges [...] [g]overnments [...] [t]o make serious commitments to restructure the global financial architecture based on principles of equity, transparency, accountability and democracy, and to balance, with the participation of civil society organizations, the monetary means to favour human endeavour and ecology, such as an alternative time-based currency.” [We the Peoples Millennium Forum Declaration and Agenda for Action: Strengthening the United Nations for the twenty-first century; United Nations document A/54/959,  adopted by the Millennium Forum on 26 May 2000; section "C. Facing the challenge of globalization: equity, justice and diversity"; this very section quoted here was contributed by John Turmel]

Choosing a time banking service or software

My interest in alternative economy is both at the theoretical and pragmatic level. On the pragmatic side, I desperately need a system to track neighborly help; because when doing that in larger-than-usual amount (say, 50 hours per month and more) and without payment of course, you better track what time you invest and what you get in return or you might eventually end up poor and exhausted.

So what I need is a time banking system. My detailed requirements, in order of importance:

• Non-local. Not restricted to any local area, neither formally nor by informal expectations or average location of members.
• Useful for a personal global network. As I will lead a nomadic lifestyle, the system should be useful within a global network of friends and acquaintances. There is no need that the system allows trade with all members of existing alternative currency systems, but that would be great of course. Trading with personal acquaintances is enough.
• Modern UI. I simply can’t bear with web-sites coming with that pre-2000 look …
• No fees. Transaction or membership fees are simply nonsense in an age of software and automation, where web-based services of similar complexity are all free.
• No-barrier registration. No requirement for an interview, for subscribing and sending in stuff by snail mail. Immediate login after registration must be possible. Pseudonymous registration must be available.
• Open source software. When installing own software, it should be free. When using a webservice, it should ideally be based on free software, too.
• Well-known programming language. I’m happy with Java and PHP and also Ruby, as this allows me to modify the software efficiently when necessary, and to better diagnose problems.
• Low maintenance. So the optimum solution would be using a web service; hosting software on an own server is possible but can be cumbersome to keep up.

Here is a short overview of the projects I found, with my evaluation. Note that I just looked into descriptions and demo versions mostly, so do not consider this my last word on this. By adequacy for the above requirements:

1. Global Groups Exchange at CES. This is the only international and time-based exchange at the Community Exchange System. Seems like the ideal choice among the CES alternatives and the only one.
2. Registering an own CES exchange. The Community Exchange System (see also in German Wikipedia) is a unique global network where local alternative currency systems like LETS and time banks can join. It provides them with a web based accounting environment, and also allows trading between the different groups (which all use their own currency). So the shortcoming of LETS and time banks of being “just local” is remedied that way! Registering an own group involves some administration efforts for that “own” group, but also allows more control and customization.
3. Registering an own time bank at Tauschen ohne Geld. It seems to be possible to create a group there that uses work time as the unit of currency. However, the system is not international (German domain, only member groups from Germany, Austria, Switzerland). Also, it is not clear so far how trade between the individual groups works; it is supposed to use the Ressourcen-Tauschring (RTR) in some way. But like CES, Tauschen ohne Geld provides an e-banking environment to members. There seem to be no fees.
4. Rivulet on own server. Ripple is a web-of-trust extension of the LETS system, avoiding the group trust necessity that keeps LETSs local. Rivulet is a Ripple server [source] written in Ruby, while the Ripple reference implementation RippleSite is based on Python [source]. RippleSite so far offers a quite simplistic UI, but Rivulet is said to support AJAX (but I did not test the features provided in the UI so far). So in all, I’d like to check Rivulet as my favorite Ripple implementation.
5. OSCurrency on own server. The OSCurrency project is a Ruby-based implementation of the OpenTransact protocol [source], specifically targeted at building a timebank. Example and demo sites etc. can be found on the OSCurrency website.
6. Cyclos on own server. Cyclos is an awesome, free and open source (!) software specially made for alternative currency banking. LETS, timebanks etc. can all be provided with a Cyclos installation, which acts as a website providing a central e-banking environment to all members. The project is very rich in features and seems to be the most mature software product in the area of alternative currencies as of 2011-08. It’s written in Java and runs as a Tomcat container, so installing can give you some trouble esp. if you already have a Tomcat with some web applications running. Be sure to check the Cyclos installation instructions.
7. RipplePay account. Ripple is a web-of-trust extension of the LETS system, avoiding the group trust necessity that keeps LETSs local. RipplePay is the portal supported by the reference Ripple implementation, RippleSite [source]. While this portal has no vibrant community yet, it could still be used together with friends if one targets trading with them only anyway. However, RipplePay’s units of currency are bound to USD parity (while of course not needing real dollars to make payments). This makes orientation simpler, but would require a hack like an internal “15 USD = 1 hour” agreement to using it for time banking. This could of course be fixed easily when installing RipplePay on ones own server; but then again, I would prefer Rivulet, see above.
8. Registering an own PicoMoney currency.  PicoMoney acts as a service provider that allows this, even for free. However, when issuing a time-based currency, the PicoMoney seems unfit to operate a LETS-type exchange network: there is only one issuer per currency, where issuing means, handing out some of the currency as an IOU (“I owe you”) statement in return for received services. This would mean that, to establish a network, everybody has to hand out an own type of currency that represents the time of just this individual. This would be still tolerable if PicoMoney would allow free conversion between these currencies, but this seems to be impossible in its current state. Note that PicoMoney provides an implementation of the OpenTransact protocol [source].
   
9. Joining zeitbank.net. They seem to accept members without respect to their current location, but their activities seem centered around Munich. Also, the site is only available in German. Also, they require a membership fee of 36 EUR for an individual, per year.
10. Karmagora. A great-looking global time-banking system. The only reason it is last in this list is, it’s still in private beta as of 2011-08.

Other alternative economy projects

That is, those I found to be interesting, but which are not directly usable or related to time banking. In no particular order:

• Kleiderkreisel. A Germany wide barter ring for clothing.
  
• OpenTransact. This is an open source protocol, with a reference implementation, that allows trading with any kind of currency, incl. “time units”.
• The MetaCurrency project. Inventing new forms of money, but seemingly in the early stages?
• A great comparison table of alternative currencies and monetary systems.
• tauschringe.de. A German information site on LETS systems, including an LETS address directory.
• Tauschbillet. A Germany-wide LETS-like system for bartering goods, with a “billet” currency as an intermediary.
• Tauschzone. Another Germany-wide LETS like system for goods, with a private currency as the intermediary.
• German LETS address directory.
• Open Money project. A wealth-acknowledging information system.

Symptoms

In my FreeMind performce problem, I had the following symtoms:

• Sometimes, reaction times from 0.8-1.2 s when selecting a different node with the mouse; after such a long-delay selection, most nodes in vicinity could be selected without any noticable delay.
• Regular latencies of 0.8-1.2 s when adding a note (by pressing “Enter”), and additional 0.8-1.2 s of latency before letters that are written into this new node appear on the screen. After that, one could write without a delay. If one would wait after adding a node until the caret starts to blink in the text field, one could then start typing in a way so that letters appear on the screen immediately, but the total latency would stay the same.
• Of course, long latencies (perhaps 10-15 s) for special operations like “select all nodes” (Ctrl+A). But this was still tolerable, as it is rarely needed.

Solution

The only solution (rather, workaround) that I found so far is to reduce the number of nodes in your mindmap. This probably means to outsource some branches to external mindmaps and link to them from your central mindmap. Depending on what you store in your mindmap, it could also mean to combine nodes into one (like transforming the subnodes of a list-type node into items of a bulleted list in one HTML node).

If your mindmap only uses HTML nodes and you do not really need HTML as you employ no formattings, convert these HTML nodes to plain text nodes. In my informal tests, this improves performance by some 30%. Using a mindmap that not even contains icons and styles seems to further speed up things (maybe another 30%), but I did not thoroughly test this.

Discussion

From my informal tests it seems that FreeMind latency times in the above mentioned cases linearily depends on the number of nodes. Much text in the mindmap is not so much of a problem: I had a 2.8 MiB mindmap with much less nodes (estimate: 1500) and could still comfortably work; even though most of these nodes were HTML nodes … .

After splitting the problematic mindmap into two (one with 6000, one with 4000 nodes) I could work with sufficient comfort again. Inserting nodes and starting to type now took 0.4-0.6 s (in the cases where high latency is experienced). Selection was alwayss instantaneous.

Also, it seems to not influence FreeMind performance with large maps if your nodes are descendants of an encrypted node (like in my above mentioned 10 000 node mindmap) or not.

• The SeaTools bootable CD booted ok, and the initial SeaTools interface showed up with a licence acceptance screen.
• Mouse was shown as installed as “PS/2″ during the FreeDOS boot.
• No mouse or keyboard input works. This also applies when booting up with a USB mouse connected.
• Ctrl+Alt+Del brings one back to the DOS level, and here the keyboard works.

This seems to be an old unfixed problem on some systems, see for example forum discussions here and here.

Workaround: this problem appears only on some systems, not on all. So I simply switched the hard disk into a different computer (ThinkPad T60), and here mouse and keyboard work.

• Good integration with LibreOffice 3.4 (practically identical to OpenOffice.org so far).
• Free and open source software. If this is impossible, it should be a web servicewith good export options, to avoid vendor lock-in.
• Collaboration on one bibliographic database by several users via web or desktop clients.
• Cross-platform compatibility (Linux, Windows, Mac).
• A project in active development.
• Public read access on the Internet (a regularly auto-exported bibliography list would be sufficient).
• Ability to attach full-text files to reference database entries; these should be accessible by co-researchers only, not the general public.
• Ability to mark favorite quotations in full-text files via bookmarks, or extract them into another file attached to the database entry. This should also include the option to add notes and other meta data like a summary to prospective quotations.
• Ability to have an online repository of full-text files and sync it to ones local system.
• Simple data entry, ideally by auto-capturing of bibliographic information from web pages.
• Various bibliographic output formats, ideally via Citation Style Language.
• The exports in various output data formats should also be accessible by the general public, like for example in the StarTides Reference Library.

Result

Up to now, I found Mendeley to be the most adequate reference manager application for the above requirements. It basically meets all of the above requirements, and as per their own comparison sheet they claim to be somewhat ahead of their competition.

However, I did not start to work with Mendeley now. When I made my firs experiences, I will report back!

Discussion

My decision for Mendeley was done after a close competition with Zotero. It is surely worth a look as well! The only set of requirements not met by Zotero is annotating PDFs and sharing these annotations with collaborators. (Also, Mendeley comes with more free storage space for papers [source]). Apart from that, it meets all of the above requirements (in some cases, with some hacking):

• The project is in active development as of 2011-08, and backed and funded by scientific organizations which lets one expect this to be an active project way into the future.
• There is a plugin for integration with OpenOffice.org / LibreOffice.
• And it is said to include a feature for “online syncing” [source].
• It supports Citation Style Language.
• Full text files can be synced across collaborators using a dropbox.com account, for example. If necessary, the ZotFile Reader extension can be used to push files into this folder automatically.
• There are also extensions for the Zotero Firefox extension. Search for “Zotero” in the Firefox extension directory, and you will find:
• ZotFile – Zozero plugin to rename, move and attach PDF files to records.
• SEASR Analytics for Zotero – Citation network analytics via SEASR.
• Zotero Scholar Citations – Citation network analytics via Google Scholar.
• Zotero autoexporting – Exports the bibliography database automatically in a desired format.
• ZotFile Reader – Comfortably read the full text of a bibliographic entry on a tablet device.
• ZoteroQuickLook – Mac OS X style “Quick Look” preview of full text files attached to bibliographic entries.
• Zotero autotranslating – No idea what this is for.
• Zotero is also recommended as the current best solution of bibliography management in OpenOffice.org, by the OpenOffice.org bibliography project [source].

Some other alternatives and why they were not chosen:

• Bibus. It is possible to do collaborative work on the bibliography by means of a shared MySQL database [source], and this software also has great integration with OpenOfice.org / LibreOffice. However it still seems to lack capturing abilities for bibliographic information on web pages, and also still seems to lack features for attaching full text documents (I hacked a solution for that when using it in 2006).
• Qiqqa. The features seem great but it is not free software. And doing more than just bibliography management in a proprietary software (in this case, note-taking and mindmapping) smells too much like vendor lock-in.
• Aigaion. It’s free software and completely web-based, which is great for shared editing and public access. However it seems to lack dedicated integration features for OpenOffice.org.
• Jumper 2.0. Really interesting software: a knowledge / recommendation engine that allows to record metadata about all kinds of objects. However, we rather need just a good software for bibliography management, including producing bibliographic references in the correct format, and Jumper seems to be weak at this.
• refbase. Interesting feature list, including full web-based environment and capturing of records from web pages. But the latest release is from 2011-08.

The best / most comprehensive sources of researching for a fitting reference management software have been:

• Wikipedia: Comparison of reference management software
• OpenOffice.org wiki: Bibliographic/Software and Standards Information

1. To be able to set up a cheap high-speed long-range computer network connection using drilled copper wire, which is the simplest and cheapest cable that can be used for this purpose. Bell wire 2×0.8 mm² on 250+ m coils costs ~0.1 EUR/m.
2. Setting up such a connectio with minimal hardware costs. Using SDSL hardware is the cheapest alternative as of 2011-07: both SDSL routers together have been obtained for aboute 40 EUR in used condition, while a similar setup with VDSL2, which of course is up to 100 MBit/s, would be 360 EUR (AllNet ALL126AM and ALL126AS) or with ADSL it would about be about 200 EUR at least (CATENA CN24 MODEL 211 MINI-DSLAM, and a simple ADSL “client-side” router). The problem is that VDSL and ADSL server hardware is nearly impossible to come by in used condition; while this is possible for SDSL hardware, as server functionality there was seemingly more common as “just another firmware feature”.

The equipment you need

1. SDSL server. In this case we use a Zyxel Prestige 681.
2. SDSL client. It is highly recommended to use the same product that you also use for the SDSL server, just configured as a client. Because, SDSL hardware that can be used as a server normally is meant to be used that way, and tested that way. Trying to use any ordinary SDSL router or modem on the client side might or might not work. In this case, using a NETOPIA R7200 SDSL router as a client for the Zyxel Prestige 681 did not work out: even though the same link compatibility mode (here: Ascend/Lucent) and speed (multiple tests with different speeds) was tried, syncing failed. If you want to try a normal SDSL modem: a modem has to be connected to a PC in order to make a connection. So in case you want to connect multiple PCs to an Internet connection on the remote node, better choose a router.
3. Telephone wire. You can use any two-core bell wire or phone wire, but note that the speed and reach of the connection depends on the cross-section. For example, using AWG 26 wire (means 0.129 mm²) one can use the full SDSL speed of 2320 kbps up to a distance of 2.3 km [source]. It’s also noted there that using  thicker wire increases this distance.
4. WLAN router (optional).

Note that SDSL devices are rare in Europe as they have never been in widespread use here, but more common in the U.S. though falling out of use there, too. For that reason, obtaining two pieces of the same SDSL hardware, as proposed here, can prove to be quite difficult. You may take some months of observing eBay auctions, and better look to the U.S. market as SDSL had been way more common there.

If you want to try combining different devices as SDSL server and client (you have been warned not to do so!), selecting your SDSL device, you also need to take care to get matching SDSL flavors: SDSL was a proprietary protocol, and thus has multiple vendor-specific SDSL flavors. The common ones found in the wild are Ascend/Lucent and Copper Mountain. In many product descriptions, these identifiers are missing, but you can still detect the supported flavors by the supported speedsteps:

• Ascend/Lucent link compatibility mode is supported if the following speeds are supported: 144K, 272K, 400K, 528K, 784K, 1168K, , 1552K, 2320K [source].
• Copper Mountain link compatibility mode is supported if the following speeds are supported: 160K, 208K, 320K, 416K, 784K, 1040K, 1568K [source].

Note that there are many devices which support both modes.

List of SDSL devices with “SDSL server” / “Central Office (CO)” functionality

• ZyXEL Prestige 681. Resources on that device:
• manual, configuration examples and firmware
• support note (the missing document in the previous source because of a dead link; includes all kinds of useful information from FAQs to password reset etc.)
• Net To Net (Paradyne/Zhone) SNE-1000 SDSL Ethernet Extender Kit (containing two devices in one package)
• some model of Speedstream SDSL modems, with support of the Ascend/Lucent SDSL flavor [source]

Setting up the back-to-back connection

In case you have the ZyXEL Prestige 681, refer to the document “LAN-LAN Konfiguration Prestige 681 (Bridgemode)” (German).You might want to make adaptations to get a routing configuration, which puts less load on the SDSL connection as no broadcast packages are transmitted.

In case you have different hardware, refer to its manual and maybe to the collection of tips below.

Tips on using SDSL

From my limited experience (note that my attempt to set up the connection with ZyXEL and Netopia devices failed!), here are some interesting notes and hints on starting to work with SDSL:

• The best source of technical information on SDSL on the Internet seems to be the OpenWAN project.
• It might be that the Netopia R7200 with the SDSL/ATM wanlet (as was mine) only supports “flavor A” SDSL and therefore could not sync with the ZyXEL Prestige 681 which supports only flavor B (namely Copper mountain and Ascend/Lucent). This seems strange as the configuration menu also offers the option “Lucent”, which normally means flavor B in variant “Ascend/Lucent”. But in the OpenWAN project they state that “[o]ur experiments using the Hack-o-Rocket have revealed that all profiles other than Nokia correspond to Flavor A” [source].
• The Netopia R7200 as SDSL router can come in two variants: with the Copper Mountain wanlet, or with the SDSL/ATM wanlet. You can differentiate between them by what link compatibility modes are offered to you in the WAN configuration: if it includes Nokia etc., it’s the SDSL/ATM wanlet. If there is no such selection at all, it’s the Copper montain wanelt. The Netopia R7200 manual for example is written for only the Copper Mountain wanlet.
• If you want to combine a SDSL server with a SDSL client that is not the same product, they need not only support the same SDSL link compatibility mode and speed (see above) but also the same encapsulation mode. Encapsulation mode refers to how MAC level network package information is “packaged” when travelling over the SDSL link. Compatibility in this matter might simply not be the case, as there is a chaotic set of different choices. In just two devices I have seen these:
• RFC 1483
• RFC 1490
• CM 1483 (which is the “Copper Mountain” SDSL flavor’s variant of RFC 1483)
• CM 1490 (which is the “Copper Mountain” SDSL flavor’s variant of RFC 1490)
• PPP (with various authentication modes that did not match either: None, PAP, CHAP, PAP/CHAP, …)
• and there are more
• It seems to be irrelevant for SDSL if you switch its two copper cables. At least my setup behaved exactly the same when doing so (starting to sync, the client immediately after the server, but failing to sync finally for reasons explained above). The reason seems to be that both cables are used for transmitting data (no “ground”) and that syncing will make the system learn which data should go in which direction. See this quote: “[A]t the bitpump level the SDSL line is not much different from one of the loops of an HDSL system. [...] An HDSL system consists of two metallic loops like a traditional T1 system. However, whereas in a traditional T1 system each loop carries data in only one direction, in an HDSL system each loop carries one half of the DS1 payload in both directions.” [source]
• At OpenWAN, there is a great technical overview (and rejection of their design choices) of Netopia R-Series SDSL routers [link].

Conclusion

SDSL is a “severely broken” technology that never was independently standardized and thus is a total mess of proprietary “standards” and their extensions and variants and flavors and modifications. The only meaningful way to operate a simple SDSL back-to-back connection (as we intended here) is to use two exactly identical devices of which one will be configured as the SDSL server and one as the client. Cross-device Interoperability in the SDSL sphere is a nightmare. The devices were just meant to be provided by one ISP to his customers, and thus the ISP only had to solve the interoperability problem with his SDSL servers once and for all (but even that is too much for a single back-to-back connection as we want).

For anything that is not a simple SDSL back-to-back connection between identical devices, better never ever touch SDSL. It’s a total mess.

Let’s close with a quote from the world’s “most total” SDSL nut

In fact the lack of any semblance of a standard for SDSL and the proliferation of wildly different flavors makes it virtually impossible to market SDSL CPE to end users. How would an average user or even someone who is an experienced network administrator but not a total SDSL nut like me figure out which CPE version or configuration to choose? [source]