Just a quick note to let you know that Websaver should now work on 10.6 64bit. The recompiled version is available here:

• websaver-2.3.dmg

As you can see, I have moved the source to Google Code, and so any issues can be reported on the issues page.

Cheesy disclaimer: I don’t have a 10.6 64bit capable machine for testing :-( . Please let me know if it works.

Bug Fixes:

• User@host:port style names should work properly now
• -x -y arguments documented properly
• More meaningful warnings if your configured clusters are badly declared

New Feature:

• “f” in bounds mode, now resizes csshX to take the full screen.

Get it here:

csshX-0.63

New features:

• Improved tiling:
  • Can specify rows/columns on the command line
  • Can change row columns Ctrl-a g and Ctrl-a G

• Prevented spawned ssh commands from polluting bash history

Get it here:

csshX-0.62.tgz

We’re back!

First I would like to thank you all for your support! I believe that your emails, posts and comments, regarding yesterdays blog post, made a difference. It’s great to hear from you – it’s what makes writing software for fun, erm.. fun!

I just received a detailed message from the Director of Products at Ordnance Survey, apologising for the communication regarding the disabling of the service. He explained, that this is not a “derived data issue“, but rather to do with the licenses:

The issue we seem to have with your application is related to the Terms and Conditions of Google Earth, which end users viewing your data accept when they decide to use that product. These Terms state that by using your content (Ordnance Survey data) in Google Earth you/they grant Google a right to that data (“By submitting, posting or displaying Your Content in the Service, you give Google a perpetual, irrevocable, worldwide, royalty-free, and non-exclusive license to reproduce, adapt, modify, translate, publicly perform, publicly display and distribute Your Content through the Service and as search results through Google Services”). You or end end users of your application do not have the right to grant Google any rights to Ordnance Survey data, and that’s really where the problem is. Google claims similar rights when you use the Google Maps API.

We at Ordnance Survey do not understand why Google would require this perpetual right to data that their users only want to display, and are therefore working with Google’s lawyers to get these terms changed or at least get a formal clarification.

IANAL, but I think this is referring to the Google Earth License and it’s reference to the Google Terms of Service. Now to me this seems to only refer to search within Google Earth. That said, I couldn’t see references to the Google ToS in the Chrome license.

In the mean time, the OS have already re-enabled my Openspace API key but:

The purpose of re-instating this key is to allow you to explore these other tools. You will need to make it very clear on your web pages that the KML stream is not to be used within Google Earth, as that would breach your licence.

For now, I will add a this to the relevant pages, but I hope we can clear it up soon.

EDIT 2009/05/19: The service has been re-enabled – see this post…

What’s happened?

After running my KML based Openspace Overlay for Google Earth service for the best part of a year, the Ordnance Survey has decided that the service is in conflict with the OpenSpace Developer Agreement, and I have had my API key disabled.

Although I have requested clarification on exactly what aspect of the agreement I am in violation of, or how I could make changes to to the service so I can comply with agreement, they have chosen that they would rather block the it. I will give full details of my conversation with the OS below.

What can you do about it?

Over the year, many of you have sent me great feedback of how you are using this service for many things including hiking, biking and even in schools. I have really appreciated this! I believe this service has benefitted the community, and that is why I have worked on it and hosted it (at my cost).

I believe the the Ordnance Surveys actions are unfair and unnecessary. If you feel similarly, here are some things you might like to do:

• Contact the Openspace team at the Ordnance Survey at let them know how you were using this service.
• Help shape the future of the Ordnance Survey by voicing your opinions about their future “strategy“.
• Follow the Free Our Data campaign to find out why we should be allowed to use our data.

Finally, if you do follow up on this, and have time, drop a comment at the end of this post to let me know!

The full details

The Ordnance Survey Overlay started life in October 2007 – it was a simple CGI script that translated data from the OS Explorer web service in KML suitable for Google Earth. In early 2008, the OS released Openspace API, and I registered for my API key and ported the service code.

From the beginning, I realised that this was going to be an edge case for the Openspace API, and I took great care to ensure that there could be no commercial conflict of interest and to maintain all copyright notices prominently. At no point did I store any OS data on my servers. I just provided users with pointers to publicly available map tile data on the OS web servers.

The service ran smoothly (with the occasional outages late in the day due to the 30,000 tile limit), until on April 17th, 2009 I received the following message from a member of the Openspace team (out of politeness I won’t post their contact details):

Regarding your project that displays Ordnance Survey data in the Google Earth application, we are writing to let you know that the OS Openspace Developer Agreement is inconsistent with what we understand to be Google’s standard terms and conditions and any provision of OS Data to Google for display on Google Maps would be in breach of the OS Openspace Developer Agreement and an infringement of Crown Copyright. Your project is therefore not currently allowed and the use of OS OpenSpace should be discontinued.

Now, since I am not using any Google APIs – I responded on April 19th:

Thank you for your feedback on my project.

The KML Overlay for Google Earth has been a service that I have been running since last autumn. The advantages of being able to take two dimensional OS map data and visualise it in 3D, or compare it to satellite imagery in Google Earth is something that has not been readily available to the general public, but is proving very useful not just to me, but to many other people. For example, I have received positive feedback from a number of users who have been using the service for planning safe routes for outdoor activities. I have also heard from teachers who are using it to explain mapping concepts to school children.

For the benefit of the community I would like to keep this service running, so I would be very grateful if you could clarify exactly where I am in breach of the OS Openspace Developer Agreement, or infringing Crown Copyright, and if so, what steps I could take to rectify this.

From my point of view, I believe:

* I am not providing any data to Google, nor am I storing it myself.

* I have included your copyright notice with the displayed data.

* I am in no way profiting from this project (in fact I am covering the costs, such as hosting and bandwidth personally)

* I am not using the Google Maps API and so I do not believe I am bound by it’s Terms of Service.

After a long silence, I finally got a response on April 29th:

We really like your project, but unfortunately it will need to be discontinued as soon as you can. This is because it remains our current position that OS OpenSpace Developers may use third party content with our data provided they do no grant the third party any wider rights than those the Developer agrees to when accepting the Ordnance Survey’s OpenSpace Developer Agreement.

Now they seemed to believe I am granting “rights” to Google. Me response (after a short vacation) on May 7th:

I’m still a little unclear on exactly what I am doing wrong. You mention that you believe that my service is granting a third party (I am assuming Google in this case) wider rights than the OS is granting me, however I don’t understand why this is the case.

My service is just translating the OpenSpace HTML link data into the similar KML format. This is then viewable through any KML browser, of which Google Earth is the most popular.  By providing a KML version of the data, complete with the OS copyright notices, I don’t believe that I am granting Google, or any third party, any rights. Isn’t the argument similar to saying that you grant Microsoft rights by viewing the data in Internet Explorer?

If the _appearance_ of granting rights is an issue, perhaps I could add further disclaimers, or copyright notices to the displayed data?

Finally, today, April 18th, I received:

Following our email on 29th April 2009 about the OS OpenSpace terms and asking you to discontinue your application we have not noticed that this has happened, so we have deactivated your account.

If you would like more information about OS OpenSpace terms please check out the website at http://openspace.ordnancesurvey.co.uk/or contact us at openspace@ordnancesurvey.co.uk.

If you find that you would like to use OS OpenSpace in the future for a new project please let us know and we will activate your API key again.

The Future

I don’t expect any quick fixes to get this service back on the air. Other people could run my code with their API keys, but I expect they will get their keys disabled too, probably quicker than mine.

I would still like to release the code – this is something I should have done ages ago. It’s a mess, but not complicated. I will put up a post up here when I do this.

There are also other ways to get at OS map data. I expect that a client side tool may solve the problem in the future.

To follow just my Google Earth posts, subscribe to the RSS feed on: http://gavcode.wordpress.com/category/maps/google-earth/

New features:

• A new “[s]end” mode that allows you to send host-names/connection strings to all windows
• ssh-options can now be configured or set on the command line.
• p in bounds mode shows you your current bounds, good for pasting into csshrc
• Ctrl-a Ctrl-m to minimise all windows including master
• Ctrl-a Ctrl-h to hide all windows and minimise master (my favourite :-))

Bug fixes:

• Infinite loop on “enable mode” when some windows have closed.
• Some document corrections

Available at: http://code.google.com/p/csshx/downloads/detail?name=csshX-0.61.tgz

Enjoy…

I just uploaded csshX  v0.60.

Main new feature is draggable bounds changing. This allows you to resize or move your window set after it has been opened. Bounds changing is accessible using Ctrl-a b. Once you select the new area for the windows to use, press Enter to accept it.

* Note, dragging between Spaces on 10.5 does not work. I cannot find a way to ask a window which space it has been moved to, nor tell another window to move to a space.

You can also now hide your window (rather than minimise them) with Ctrl-a h. This is much faster than minimise, and I find Ctrl-a h, Apple-m is a great way to clear my csshX screens away temporarily. Window hiding has also allowed me to improve the retiling performance.

Other changes are minor, including one fix to the nasty infinite loop bug when using enable mode selection.

As always, please report any issues you may come across, preferably on the issues page.

Thanks to some great publicity from Mac OS X Hints, I have received lots of feedback and bug reports.

Number one issue has been multi-monitor support, and problems related to it. This is now working.

To use it, add a –screen n to your command line, where n is the screen number. It will default to screen zero if omitted.

This was actually a nasty problem since there is no quick AppleScript solution, and the only way to get screen counts and sizes is by calling the AppKit frameworks. This is Objective-C only and since csshX is written in perl, I had to mess with the PerlObjCBridge (or Foundation.pm as it’s known) to get the NSFrame. To make matters worse, NSFrame is not even Objective C, rather it’s a pure C struct. In the end, I use NSValue to give me a string “description” of the data, and parse this with regex’s (differenet for 10.4 and 10.5 naturally!).

Other changes:

• Cursor keys now working for terminal applications like more and vim
• Windows will no longer cover the dock (wherever it happens to be)
• A –debug option for when things go wrong
• Some documentation corrections.

The package is available at:

http://csshx.googlecode.com/files/csshX-0.59.tgz

For some unknown reason, my OS OpenSpace tile limit has suddenly dropped form 30,000 to 1,000. As you can imagine, this pretty much cripples my Google Earth overlay.

Thanks to everyone for pointing it out – it’s great to hear that people are actually using this!!

I’m looking into it, and will update this post when fixed…

One of the all time great sysadmin tools is Cluster SSH. It allows you to open up ssh sessions on multiple machines and send commands to them all through a common master window. It’s a perl script that uses xTerms and Tk.

My problem was that OS-X has a bit half-hearted X11 support and I always wanted a version that used Terminal.app. Finally after messing with some Applescript the other day, I realised it is possible to do.

So I have created csshX – my OS-X homage to cssh.

csshX is pure Perl (with calls to AppleScript) and should run on any 10.4 or 10.5 machine.

• Download
• Manual Page

One known issue is that if you change focus, while csshX is opening many windows, it may capture the wrong terminal window. This is because there seems to be no good way to capture a handle to a Terminal window while you are opening it (I have to grab a handle to the front-most window after it has opened).

Let me know if you find csshX usefull, or you have any good ideas about how I could improve it.