Well, except that Access is not well-suited for typical pipeline use.

My objection is that the article is not Linux specific. The same program can probably be installed easily on OS X or FreeBSD.

"The Joy of people who make their software available" would be a more appropriate title.

No need to shit on the guy.

TFA boils down to: "I was emailed a proprietary format for an application that doesn't exist on my platform. I was able to find and install a third-party utility to convert it to something usable."

An astonishing tale of Linux superiority it's not, but I wouldn't say they jumped through hoops or spent hours researching anything here.

  > [..] we managed to do something that is 
  > absolutely trivial for everybody else [...]

Where everybody else means anyone with a copy of Microsoft Access. Take it for what it is: guy's excited that he was able to use free tools to perform a task 1) that he thought would be much harder, and 2) potentially faster than his colleagues with the official tool. So what his conclusion's a bit of a leap?

He managed to find free, command-line tools to read a proprietary data format. This means he could potentially wrap a simple script around them and automate the process if he needed to do it on a regular basis.

Would this be free & scriptable on Windows? It's been a while since I had to do anything comparable on Win, but I think he would have had to buy MS Office (Pro?) with VB or Visual C++ to do it.

It would be completely trivial on Windows since mdb files are accessed with the Jet database engine, one would set up a database connection string just like when accessing an Oracle or SQL Server database through ODBC and then use whatever database api is available in one's language of choice (vbscript, javascript (both of them available as 'standard' cscript engines on any installation of Windows since 2001), powershell, visual basic, C++, ...) to extract the data.

Or, one would google for 10 seconds like I just did and find for example http://timothychenallen.blogspot.com/2007/08/vb-convert-any-... which does that for you and converts the file into a csv file, and then process it like other csv files. No Office or Visual Basic required.

So his claim 'this would have taken a couple of hours' (about converting the mdb to csv on Windows) is a load of bollocks and FUD. He didn't even search for it I'd wager, and let his ideologically myopic view cloud his pragmatic judgement. I'm not sure why people are defending this sort of juvenile fanboyism, it's drivel like this that makes the rest of us serious users who propose Linux in situations where it actually is better than the alternatives look like nutcases.

All the dev tools (C++,C#,VB,ASP etc.) are free on Windows via Visual Studio Express. And although I've never tried I'm pretty sure the database drivers needed to read and write .mdb files are included with Visual Studio Express. So no need to buy anything.

You could probably script it with Powershell and a bit of .NET

mdb files are not easy to open, unless you have MS Access. Lots of Windows computers lack MS Access.

Now, you can still open mdb files on Windows, using the Jet Engine and something like ADO, ADO.net or ODBC, but the details of that are a bit too gory for the mild mannered command line junkies here.

Sure, and for a reason - they contain a lot more information than a csv file can. It contains metadata on the table columns, relationships, etc. We don't complain that MySql or Postgresql files are binary, so using 'it's binary' as an argument against mdb is bollocks.

Of course, maybe the source of the data was wrong in choosing mdb as the format, and a csv would have sufficed. I don't know, that's not the point here. The point is that mdb is the easiest way to get 'database-like' information that is more complex that just a flat list of records into the hands of 'regular' users, as much as that may dismay those for whose platforms nobody has written software to open them.

You are correct in that a database file may contain more rich data that CSV. I think it is open to debate, however, whether a closed format like MDB is easier for the end user, over something open like SQLite.

I'd estimate that 90% of the Windows computers out there that are used for 'everyday' work have Access on them. For those users, opening the file is basically double-clicking it.

For that to work with sqlite databases, one still needs to install a GUI tool first, hope that it associates the right extension with the sqlite file you have, and then get over the database-centric user interface that all of them have.

Look, I'm not saying that mdb is better than sqlite. But I am saying that the notion that the 'openness' of sqlite somehow makes it easier for regular users to open them is a crock.

Or maybe we can talk about if using sqlite databases in simple, 'script-like' applications is easier than doing the same with mdb/Access databases. That is more nuanced and depends on what environment you want to talk from etc. But that goes far beyond the definition of a 'user' as in the use case of the OP (a large student body to whom exam results need to be communicated).

I would estimate the number of Windows computers used for "everyday work" which have MS Acces on them is more like 15%. That is just in my personal experience though.

OK fair enough, I don't have anything more substantial than a wild-assed guess to back up my 90% claim. I'd still say that an mdb file is closer to the everyday experience of the average Windows user than an sqlite file is. And even if they're equally difficult to open, the 'openness' of the sqlite data format is still of no importance whatsoever to the average user, which is the original claim I was opposing.

I think this is primarily about finding a hack which is a short-cut to doing something in the officially recommended way (with Windows and Access)

> And the resulting way is far simpler and didn't require paying for anything.

because, of course, time is free.

Exporting the file to a csv format (using microsoft access) wouldn't have taken any less time than me finding the right tool on linux and doing it, that is if I had Microsoft Access.

And the resulting way is far simpler and didn't require paying for anything.

How about this story? "Something goes wrong and shit doesn't work. At this point, we would expect to be completely stymied. But no, instead, we manage with a bit of googling and dicking around to fix it!" The comparison isn't with opening this particular file on different operating systems; it's with how you can recover when shit goes wrong (which does happen now and then on all operating systems).

From that story one doesn't draw the conclusion 'Linux is awesome'. At best, it would be 'Google is so cool, I allows me to fix things'.

Let's face it, this 'article' was utter fanboyism, and does not exhibit any form of proof that Linux is better than whatever.

(I've run Linux since 1998 and I do prefer it for some tasks, like server tasks. I have many years of experience and am well aware of its strengths and weaknesses. Interoperability with the files that 'regular' computer users use is not one of those strengths).

Yes.

I recollect a similar situation when I volunteered for a Linux fest. (http://wiki.sampada.net/GNU_Linux_Install_Fest) The venue could only accommodate 50 odd people but > 300 had registered on the eve. We needed to update all that after 10:30 am seats may be full and we might not be able to accommodate all, at 10 pm.

Me and another friend of mine wrote a simple script to fetch mobile numbers of registrants, and another silly one to use gnokii and send an SMS to the fetched numbers. (http://wiki.gnokii.org/index.php/User%27s_Guide#SMS_Sending_...)

Linux is fun.

Again, not an example of any form of superiority of Linux. 30 seconds of google shows us http://www.oreillynet.com/wireless/2003/10/10/sms.html , allowing you to (with a free tool from Microsoft) do the same. Oh, and you can do it over Bluetooth and infrared, in 2003.

Have you ever tried to get infrared data connections to work on Linux? I have, and I can tell you that it takes many, many times longer on Linux just to figure out how it should work (in my case, infinitely longer, since I never did get it to work).

I'm not here to say that one is better than the other, just that the examples here do not prove one or the other, or not even give a hint in the direction. They're all completely trivial things that can be done by a competent user on any OS.

Not to nitpick, but the above linked post says that the author of Windows XP Unwired too could not get smssender.exe to work in command line mode, to send messages in a batch.

>> his argument is basically 'Linux couldn't open a file that everybody else could, but by dicking around with random half-assed documented tools we could do it anyway!'.

May be one could even interpret it like 'there are a lot of software vendors who force their users into proprietary solutions and formats but with Linux at hand, limitations could be worked around to get things done'.

Perspective, isn't it?

No, one couldn't, or at least not from the situation presented here.

Look, there are two situations in the example in the OP:

- 1. User has 'regular' Windows PC with Office. User downloads file from website, double-clicks file, is on his merry way.

- 2. User has Linux PC and an ideological axe to grind so doesn't want to use anything else. User needs to hunt internet for software, turns out not be be so easy, spends hours and finds intense satisfaction in accomplishing something that is considered trivial by other people. User then proceeds to write blog post proclaiming, totally and utterly non sequitur, superiority of Linux because he managed to accomplish said trivial task.

The only 'limitation' here is a self-imposed one - the guy refuses to use anything but Linux to convert his data on. That's like saying 'I ran a marathon with a rope of 30 cm tied between my ankles'. Impressive (in a certain way), sure - but why? Why not just run the more efficient way like everybody else, with no rope?

If you're going to defend your OS by cherry-picking tasks that are somehow easier or more elegantly done by it, go ahead - the internet is full of them, but hey what do I care, everybody is free to do whatever they want in this regard. But this 'example' doesn't make any sense and confirms the general notion that Linux is for people who like technology for technology's sake and revel in finding convoluted, over-engineered solutions to the most menial tasks. And in that sense, it affects us normal people who just want to get stuff done too, because it causes a negative image for Linux, rather than what I'm sure the author thought would be a positive one; hindering its uptake in contexts where it is superior or appropriate.

I'm not saying this guy single-handedly destroyed Linux, I'm just really desperate with seeing this toxic single-mindedness still being so pervasive, after all those years.

I don't see how this has anything to do with "Linux" - if anything, it shows how powerful google/the internet and the open source community is since nowadays you can expect to be spoon-fed a solution for almost every problem you can come up with.

And I think "Linux IS great!" articles at this day and age are pretty much unnecessary - this battle has been fought (rather) long ago.