MacInTouch Reader Reports: ODBC and the Mac covers a couple years of struggle in Mac-land regarding the availability of ODBC drivers for providing distributed database connectivity.

I’m looking into it because I’d like to read-write to my MySQL implementation from Word, not someting I expected to succeed at since MySQL is a direct threat to SQL Server’s market.

To my surprise, it’s not impossible, only tough. First, I still use Word98 in Classic, which does not provide easy ODBC support, although you can make it work if you twist your body around into a sufficient number of uncomfortable positions.

However, in both Word 2001 and Office v.X, mac ODBC is much more readily supported via a number of routes – none super user friendly, I think; but less of a forced march. Mind you, this is some connecting of the dots going on here – I did NOT find a perfect example of a simple, user oriented how-to anyplace, except maybe here:
AugSOFT Office98/2001 docs.

The problems with these older drivers are twofold: one, they’re not free, natch; and two, they don;t offer MySQL support, since MySQL had yet to gather steam back then.

AugSOFT offers a standalone server-bridge (for wintel) that can work with MySQL, but that’s beyond my scope.

On the looks promising front, www.iodbc.org and Openlink Software both offer work in the direction of tying MySQL to ODBC connectivity; as seems to be the case for most of this, it looks as though nothing is perfectly meeting my Word-MySQL jones on the matter.

hmmm.

A PHP web GUI dev studio called DaDaBik. (Psst, I’m referencing the the previous entry… sorry for the lack of context.)

Um, sort of.

I guess the way it would work here is:

1. design schema
2. create tables (CLI, DaDaBik, MySQLAdmin)
3. poke at GUI (hand, DaDaBik, Dreamweaver)
4. repeat

Geez, that doesn’t sound real elegant.

Hm, there seems to be a family of these out there:

PHPLens makes some interesting claims for itself. Oooh! Image upload. <Homer voice>ohhh, im-age up-load</Homer voice>

Oh, nm. DaDaBik is GPL; PHPLens is mucho dinero.

Take part in an exciting and life-enhancing fat-chewing session! Waste valuable time pontificating on thinly-supported opinions!

You may recall my noting that shortly I’ll be performing regular news gathering, mastication, and regurgitation duties for Cinescape, which will require cruising entertainment-industry news sites for valuable lies about upcoming projects et cetera.

Much like speeding through my blogroll, actually. But now I’ll have to pay a bit more attention. Of course, those of you that took longer than most of us to learn how and why one’s shoes must be tied, the importance of good hygiene in attracting a mate, and why it’s better for your monkey-status to shoot some hoops than read a book are thinking: “Hm, I could whip up a robot that could do that in no time flat.”

Well, first, that’s what I’m thinkin’, too, beanie-bros; and second, we’re lying to ourselves.

It will actually take a significant amount of effort.

On the web, as you well know, “a significant amount of effort” means first, we design a database. (Then we screw up and lose all the data, or maybe even the data-structure, but we don’t tell anyone and it’s back together again by morning, maybe even with most of the data)

Then we adjust the design structure to reflect the actual needs of the users (well, ok, the budget doen’t actually include this, but it says right here that we should – it’s those darn suits, holding everybody down again), since we were only imperfectly able to anticipate the data structure needs.

NOW we can work on the robot.

But guess what? That whole swath of mildly amusing geekrant is just bait. Really, it’s like this.

I have to store a great deal of data, daily, for rework and attribution and so forth. There’s a decent CMS (content management system) at the mag which I’ll be using, but they want the finished product, not the raw materials. So that means I need a locally maintained database that allows me to store articles, attribute them (via URL at minimum, but possibly with greater granularity), quickly prioritize and sort, and rewrite. I can keep in mind we’ll eventually be building a little seeker-bot to grab the data as I design the storage structure.

So here’s the question: I have and use FileMaker Pro (uh, I think I’m using an older version – 4.mumble) for most of my quick and dirty database stuff, because of its’ ease of setup and UI rework. Sadly, it’s stuck in OS9 and I’m running it via the Classic environment; which really hasn’t hurt it’s core performance, but for stuff like browser-to-field drag and drop, it’s not ideal. It works fine, though.

See Jane Code! is a small visual dev studio app for PHP and MySQL. PHP and MySQL are coding langidges (and a database server) that help to run bellerophon, the overtaxed, resurrected-from-the-grave G3 laptop which serves these pages to you, my teeming tens and dozens.

Using this tool, or something similar, I can get at least 30% of the design flexibility of FMPro and have another webside app to wave in the face of oh, those many recruiters who so ignore my weekly applications for their jobs.

I prefer an organic, UI-oriented design practice, and when the client is ME, I can indulge myself.

Here’s the turnip soup: Should I stick with old FMPro 4 and take advantage of that flexibility while I fine tune my needs, at the expense of a possibly cool demo project? Or commit to an experiment and set up my data-collector as a web-oriented app from the get go?

I could use FMPro on an interim basis, you know, to get the benefit of the organic evolution; but the interface quirks of going from browser to FMPro are not conducive to data granularity.

I could also buy a new OSX-compatible version of FMPro.

Speak, you nerdly persons in your glasses! Speak! Indulge us with your deep thoughts on the matter!

Crap. I misremembered the Newton version I have – it’s a 120, not a 130. Unfortunately, that means that NOS 2.1 won’t work on it, so no WiFi driver either. 2000’s and 2100’s only, thank you very much.

I know someone who hoards old computer crap as badly as I do. Perhaps he’d be interested in a couple of 100mb SCSI Zip drives? Or even a lovely APS-enclosure 1gb Jaz? Hmmm? Sorry, got rid of the SE, finally.

I do have a nice Power Computing 604, though. Easily upgradeable to a g3 or g4. It’s just the ticket.

SquirrelMail on OSX.

This guy’s setup is essentially the same as mine. I’ve avoided php.ini to date and hope to in future, though. Use of httpd.conf server directives is preferred, according to apache.org, so that’s what I have been trying to do.

However my use of Liyanage’s precompiled PHP module is officially making me nervous, since I don’t know all of his steps in cooking it. As he’s moving on to Jagwire and I’m staying put for the foreseeable future, I must needs hit the weights again.

A bit more musing on this, I think.

the squirrelmail site makes me a bit nervous too – ALL their docs are posted as open wiki, which means that, for example, their Requirements page can be, and apparently has been, edited by anyone, including me. In the non-english sections of the reqs, a variety of imap servers, including uw-imap, are specified as valid. Yet, in English, it appears that someone left the start of that line out somehow.

Since our correspondent at bayswaterfarm is in fact using uw-imap with squirrelmail, I must conclude that the requirements page listing is in error, which then leads me to wonder about other docs on the squirrelmail site.

Horde: System Capabilities Test

hmph. wotta pain. well, this is all about the learning, innit.

As I was setting up the Ken Goldstein Project I decided that I would set up sendmail as well. Sendmail has a reputation for being the most difficult back-end e-mail program available; difficult to configure, that is. So far my experience bears this out. For me and possibly other OS X users as well the most difficult part of setting sendmail up was determining how the system identifies its own hostname global.

The way I and others solved the problem of local host host name assignment has been to manually edit the file “/etc/hostconfig” to substitute the value manually. This is unsatisfying but it works (otherwise in that section of hostconfig there’s a directive which prompts the system to, ultimately, determine its’ hostname from the first successful DNS query it makes – which can lead to interesting results depending on whose DNS you’ve opted for). Additionally because sendmail offers so many configuration and build options it can be very challenging to correctly identify where the program obtains such data as usernames or aliases. But wait!
I promised I wouldn’t bore you with sendmail details. So that’s enough of this. But I’ve even been dreaming about it, so it’s hard to write about anything else.

Speaking of writing, Cinescape has asked me to accept a regular freelance position which I am looking forward to accepting. It’s kind of an on-line entertainment news trolling position. Somehow I suspect I am about to become a huge fan of RSS feeds.

Not that I expect studio publicity websites to provide such useful tools. Hmm. Perhaps there is a business here. Journalism people, that’s nontraditional thinking, isn’t it?

One more geek note: I used ViaVoice to input this entry, and it took a long time, presumably associated with a) my needing to learn how to use the voicerec more efficiently; b) also with the program not having been sufficiently trained up yet; and c) the inherent dictation challenge of using words like “/etc/hostconfig/” . After a bit I’ll have a better feel for it. I felt like the copy was choppy, with a fair amount of needless repetition, presumably a consequence of attempting to correct non sequitirs on the fly instead of after completing the copy block as a whole.

Day two of the “expand your knowledge of BSD considered as a superset of the Mac OS self-guided seminar and cursing tutorial” continues.

Progress is being made!

Brain cells are being bludgeoned! Errors in books are being noted! The simple beauty of Mac OS Classic based mailservers is suddenly much appreciated!

Ah, but the payoff will be tomorrow, now shan’t it me hearties (actually that’s done already – I just got ambitious and trekked on into sendmail territory). And yes, Ima call it the art for the week.

I’m busting ass setting up a website using the popular, if yet a work in progress, CMS postNuke, and oog, I’m experiencing the downside of opensource collaboratively developed projects.

Which is not to say that so far I haven’t been pleased by the capabilities and featureset of the application. It is to say I’ve been quite frustrated by poor documentation, aggravating, developer-centric assumptions about how users and administrators will interact with the system once it’s in place, and partially-implemented functionality.

I opted for postnuke over phpnuke based on security concerns (phpnuke explicitly requires “register_globals” to be on in php, which is deprecated in the current release of php), and I’m, frankly, wondering if I made the right choice.

I’ll continue with the implementation under postnuke for now, but the project is apparently in extreme disarray following the motorcycle-accident death of its’ lead developer this summer. A variety of long-term developers appear to have forked from postnuke to a new project, as yet in gestation, and the rump team have released a version of postnuke with many housekeeping issues, while at the same time updating the postnuke website in such a way that older user discussions and documentation is effectively inaccessible.

Shhh… no-one tell Microsoft marketing about this.

1. go to eBay and locate a vendor selling remaindered copies of “IBM ViaVoice for Mac Enhanced Edition 2.01”. I found mine from this vendor, who appeared sufficently reliable.

2. Be sure the item for sale is a) unopened and therefore unregistered, and b) is the full retail version of the software, which comes packaged with a USB headset mic (the Andrea USB NC-7100, in my box at least). Ask the vendor a question if ncessary.

You may be able to substitute another mic if you’re already equipped.

3. Bid or use “buy it now”. sknetstore’s buy it now price is $60, but when I bid there was no crush of bidders, so I got it for the opening price of $49 + S&H.

4. Once you’ve received your product, go to IBM’s ViaVoice website, and head over to the Enhanced Edition rebate page.

5. Take stock of the choices: download the whole thing for free (200mb, not for dialup users!), pay $5.95 for a free CD-ROM to be shipped to you, or pay $20 to have them send you a full – if slim – documentation and software package. The installer you’ll get with all of these options requires the presence of the a valid Enhanced Edition CD-ROM mounted on your machine.

That’s it! I opted for the free download. Total cost to me? About $50. Total savings compared to a list-price purchase? About $150.

Please note: it’s very likely that IBM will act to close this loophole once it’s well-known, or once they conclude that they’ve serviced the whole user-base of Enhanced Edition adopters. Additionally, it should be noted that ViaVoice for X is only supported for use on “Mac OSX 10.1”, according to IBM’s support notes; it is working for me under 10.1.2; and anecdotal evidence has been proffered of it’s operation under Jagwire 10.2. YMMV, natch; but DANG! Wotta deal!