Although I have only occasionally used MetroFi's free ad-supported wifi service (only rarely can I pick it up strongly enough), I am subscribed to their mailing list. Today I got a message from which says:

n January 2008, MetroFi will replace the existing banner ads, sponsor logos and paid search features that currently support MetroFi-Free in Portland, and instead ask users with MSN SideGuide compatible operating systems and browsers to download and begin using MSN SideGuide to utilize MetroFi’s free, wireless Internet service.

o MSN SideGuide compatible operating systems include: Windows XPSP2 and Windows Vista. Windows or XP users who do not have these versions will need to upgrade first to enjoy MSN SideGuide with MetroFi-Free.

Great - so unless you have a Microsoft Operating System, you apparently can no longer use this free service provided by contract with the City of Portland. To hell with you, Mac and Linux (and *BSD) users.

I'm really upset about this bowing to the Microsoft monopoly. I think the City should either force MetroFi to allow all computer users access to the free service (just let 'em surf ad-free if their computer wont' support MSN) or terminate their contract with MetroFi and let someone else do it. This is bullshit. (And guess what? I'm a Microsoft stockholder, too! Well, barely.)

(On the outside chance that I'm wrong and this doesn't mean non-Microsoft users are SOL...never mind! But it sure sounds that way. We'll find out in January.)

Andrew

The city hasn't paid one dime for this service, and I doubt they will find another company willing to provide wifi for free.

Here's the rest of it, that Andrew did not include:

For network users who do not wish to view advertising, MetroFi will continue to offer an ad-free version of its municipal Wi-Fi service for $19.95/mo. For the minority of MetroFi-Free users without MSN SideGuide compatible operating systems, MetroFi will continue to deliver their free wireless Internet access sponsored through other ad formats. The implementation of MSN SideGuide on MetroFi networks does not impact MetroFi’s delivery of public safety networks and subscription-based municipal applications to its government customers.

What about Windows Mobile or CE, are they SOL as well?

OK, Radioblogman, sounds like I am wrong (I missed that part of the release - it wasn't on their FAQ page), and it's good that I'm wrong. I still don't like having to use something goofy like "MSN SideGuide" but since I rarely use the service I don't care. Seems like another reason to switch to Linux or Mac, just to avoid it!

Andrew

http://www.basement.org/archives/2006/01/i_dont_hate_macs_i_just_hate_m.html

That should say it all. Andrew so help me if you are a mac fanboy I may have to take you out back and commit an Old Yeller.

I personally won't use anything that requires some goofy client be a part of the connection process.

Most all of the time, this is an artificial value add. Real value adds are fine.

At least they are allowing other OSes on the network! It's not all bad in that respect.

Still, there is this artificial value add going on, and it still sucks.

Actually I'm a Linux "fanboy!" I personally can't figure out Mac OS - find it way too confusing. :-)

Ubuntu Linux 7.10 "Gutsy Gibbons" rocks! Installed it on a flash card on my laptop (soon, the hard drive). Installed perfectly the first time, including wireless card. Didn't need to load anything, not a single driver. Try that with Windows!

Andrew

Mac's now run a version of Linux.

I personally hate Linux! I use several Linux servers, for budgetery reason, a HP server running Suse Linux is about 6,000 buying a HP server running HP-UX, (unix) is about 25,000.

You can't run a companies ERP system on a free OS, the auditors would have a fit.

The issue described in this thread is the problem with free services; with free services, you are completely at the mercy of the provider. What recourse do you have? Cancel your "subscription"? Bandwidth has value, and people should be paying for it if they expect to have a reliable connection. Someday I'm sure city governments will provide all with free access, but even then the network would be maintained with tax dollars, or some other form of revenue.

For a few billion, you could add access to wi-fi across the country, along all the major highways etc. That also gives you a way to track everyone across the country, tie all the current web cams together and add bunch and intstant monitoring of about everyone.

Remember when cell phones first came out, and only worked in certain areas, and it has now spread pretty much across the country and world? Wi fi, or something similar will follow suit, if someone can come up with a good business plan, or the government finds it to its advantage to do it.

IMHO, there are two issues.

One is free and reliable, and that's tied to revenue.

The other is standards compliance and how that impacts potential users and or builds lock-in.

Anything done to improve the former, without impacting the latter, is likely necessary and good.

Nwokie, depends on the auditors, and the company authoring the software. On some ERP systems, the database is decoupled from the business logic, with the intent of using OPEN OSes to both save money and be able to scale, with few restrictions.

For mid-sized companies, this can be a very powerful business motivator. Cost per user is growing to be a pretty big deal these days as most proprietary OSes want per user connect license fees of some kind or other.

Those are annual too, and when added together, often can negate the business gains realized from using the thing in the first place.

Linux is not a free OS, it's an open one. There is a pretty big difference. If you are running SUSE, for example, it's totally supported in the same fashion as a closed OS is. From an auditors perspective, there is no difference.

Now, a BSA style auditor, whose interest really is generating and preserving annual license revenue, is gonna have lots of FUD to spread about that. A business auditor wouldn't, given the support to address risk is in place as it should be.

FYI, there are ERP systems now, running on open OSes, themselves being open. They are not yet at the level seen in the closed ones, nor do they feature the level of integration seen there either.

On the plus side, they do run across multiple OSes, and do not feature per user annual fees. User computers then, can be most anything and be useful. There is a significant cost savings in that.

Depending on one's business, these are rapidly growing into potent options.

I recently did an ROI study on integrated systems, vs, non-integrated ones. When annual fees are balanced against some additional transaction time consumed, per user, it's not always a wash, all things considered.

Yes there are, but SAP and BAAN aren't among them, neither is Oracle's ERP.

Auditors get very nerveous, if you can't show them a support contract, guaranteeing system performance.

"Remember when cell phones first came out, and only worked in certain areas, and it has now spread pretty much across the country and world? Wi fi, or something similar will follow suit..."

This is true, but it's taking longer. The era you refer to where cell service was spotty (and expensive) was only about 10 or 12 years ago. That's also about the time the internet started to soar in popularity. Yet, cell phones work almost everywhere will people live or travel, but wifi does not.

Thats because there is hugh profits in cell phones, no one has come up with a good business plan, to make wi fi (or something like it) widly available. For those that need it, and are willing to pay a premium, you can get an internet card to go in your laptop, to allow access about anywhere.

Funny thing about those support contracts.

My experience has been a dedicated system, tuned and configured, for a given task, will continue to do that task for a very long time, barring hardware failure.

Hardware failure is easily addressed and there are lots of options, ranging from very expensive hot-swappable redundant systems, to just having the necessary backup components on hand.

Over the 10 years or so, I've been involved with large computing systems, the ones I've seen fail, are those that are tinkered with. That means automatic updates, service packs, lots of software revisions, etc...

Open systems and open standards mitigate a whole lot of that crap, and are considerably more robust as a result of it. IMHO, this is perhaps the biggest failure of larger scale integration. It just does not pay off in the longer term, as there are too many interdependencies. Those dependencies are leveraged, quite often, for revenue, without adding significant value in return as well.

Of course, that's what the majority of people are into these days, so that's what gets done.

I love to point out the case of open data standards though. There are still a few of the very expensive workstations, I sold and delivered in the mid-90's, in service and operating perfectly. They were expensive as hell, but they have paid off their initial cost many times over.

These were dedicated machines, configured for a single purpose. Users access them via the X window system, on what ever they like. (as long as it can run an X server, that is)

I check in a few times per year, via SSH, over a connection that is turned on for that purpose, then turned off, on a regular basis. A coupla hours of admin per year, keeps things running nicely. There are lots of updates, but the systems running as they are, are feature complete and complete their tasks in a very reasonable time frame.

Sometime in early 2000, most people running these things have moved to integrated systems. They've not kept their cost advantage, and lost robustness as well.

Those who didn't move, pretty much are not going to, and are looking at the open Linux as their upgrade path, waiting for solid ports to arrive. Those are on the table now, this year, so we are likely to migrate things over, then settle in for another very long run.

There are a coupla enablers for this:

1. No client server use of the software. Everything runs on the box, and is accessed terminal style, via X. X does 3D over the wire, and it's very good for most use cases.

2. Data connections are done via open, or mostly open standards. (SAMBA, NFS)

3. The data itself is either proprietary for the application, and it being CAD, this is just fine. Data exchanged or published outside that environment is open. (IGES, STEP, PDF, ASCII, etc...)

4. Integration is done via script and is largely a factor of how the target user environment is setup. Exports are always in the same place, with names, keyed to times and dates, as well as part / job numbers. On the client side, the X window system is scripted to launch from the menu, shared data repositories are short cuts, and a brief workflow introduction replaces application level integration.

There is some loss of peak efficency, in terms of number of steps required to complete a task. However, the costs associated with reaching that peak, deliver a negative return, meaning it's just more cost effective to make the workflow simple and consistent, and have it leverage the users default environment, whatever that may be, to the fullest.

Admin then, involves establishing a configuration, once, for each kind of user enviornment (linux, mac, unix, win32), then deploying it when a new user is required. That's client side and can be done without having to visit the workstation.

On the server side, the scripting is done, meaning only new users need to be added. A nice bonus is to just edit the default system configuration source files, to build the environment necessary for that user to be functional, given they are added with the default system tools.

It's literally, just accessing either a web page, starting an SSH session, with either X forwarding for the GUI admin tools, or command line for those same tools. Ask for a user, you get one, and are done, for the most part.

It's an extremely powerful way to compute, from the admin and support standpoint. Users do not appear to mind, given things are clean and clear.

Must better to pay for that, have it done and functional for a long time, than it is to wait for vendors to sort it out, only to have them sort it out again to justify ongoing annual revenue that may or may not deliver value in return for the overall cost of deployment.

Where ERP is concerned, application server style deployment of this kind, can realize very significant cost savings in terms of client side administration alone. The trade-off is some lack of integration, but that also comes with a lack of dependancies too, which save a lot of time and hassle.

Given some tasks are core to the business, it does not make a whole lot of sense to entangle them with other applications, particularly if said application does not require much on the client side to perform it's task.

One other advantage of this scheme is built-in data management. When running applications via the X window system, users are in a state where they cannot actually access the data at the file system level, unless they are doing so from the application!

Have that application run as the owner of the data, via the suid function found in UNIXes, and you've a scenario where overall data ingetrety is very high, for very little overall cost.

eg: user bob, runs application owned by user data. Bob has no permission to even read the data store, but the application, running as data does, meaning a file, open operation, combined with symbolic links to abstract the actual data repository, presents user bob with selections to choose from that are otherwise invisible at any other time.

The current trend is to couple a database and complex name hashing to obsfucate the data, in the hopes a user won't actually go mucking around. Or, data is piped over the network, to be manupulated client side, only to be piped back.

The former method is very robust, but fairly suceptable to user meddling. The latter eliminates meddling, but then requires complex caching, etc... to maintain known data states.

Running stuff on the dedicated machines, application server style, really boils down to display bandwidth and elementary network issues, leaving the state of the data as a much smaller problem than would otherwise be the case with client server.

Also, complex software update tasks really only need to be performed, once on the application server machine(s). Users use the same access means as they did before, and can simply focus on new feature / functionality changes, which can be tested and published in advance of their deployment change.

Oh well, maybe we will come back full circle some day, after enough people get burned enough of the time to actually matter.

Nwokie: Mac's now run a version of Linux.

No, you are confusing Linux with FreeBSD, which is entirely different from Linux. Both are free derivations of the UNIX operating system, but they are developed by entirely different groups of people. BSD is much more tightly controlled. Security experts often prefer BSD, in part because not just anyone in the world can contribute to the code base; changes are vetted much more carefully. A real programmer could explain the real differences between Linux and BSD in the way a mechanic can explain the difference between a diesel and gas engine.

My web server runs on FreeBSD. It's a little less user-friendly than Linux, but FreeBSD totally rocks. Rock-solid. My server has been running almost continuously for more than two years (I shut it off briefly only to move it, replace some fans, and swap in a new hard drive, even though it was still working fine). I'd like to see a Windows server that can run for 654 days without a reboot.

I already mentioned how I installed the new Ubuntu on my laptop without having to install a single driver - it was 100% automatic. Zero reboots were required in the installation. This would be absolutely impossible with Windows.

Andrew

BSD is the shit!

Totally agreed on how good that OS is. One thing I absolutely love is being able to just configure the few things necessary for the tasks required.

Here's a really cool application of FreeBSD:

http://m0n0.ch/wall/

This is a free firewall you can install on an old computer or, better yet, on embedded hardware, running off a flash card. The image of the FreeBSD OS is so tiny that it can actually be embedded like this - not practical with Linux.

As for Windows...ever use those self-checkout kiosks at places like Fred Meyer? The other night I used one and as I was about to pay it...Blue Screened!!! Yep, a real BSOD. The kiosks are based on Windows 2000. I've seen them reboot before, too. Windows 2000 is probably Microsoft's best OS, but it's still not as solid as Linux or FreeBSD.

Andrew