May 20, 2000
10:20 PM   Subscribe

Camworld points to the [second page of] the Encyclopedia Brown piece (like everyone *else* who didn't actually *read* it; *Busted!* :-) and also to Comma Separated Values Considered Harmful, a treatise that the vertical bar character ("|") is a better field separator in tabular text data files than the comma or even the tab.

Is that really news? It wasn't to *me*...
posted by baylink (13 comments total)
 
A separate topic, near and dear to Brennan's heart, will be the choice of "Considered Harmful" for the title of the paper.

If you don't get it, check the bottom of his page.
posted by baylink at 10:21 PM on May 20, 2000


And, *damn*, Bren; you're right; she's gorgeous, isn't she?

I'm most of the way into stage 5 by now...
posted by baylink at 10:23 PM on May 20, 2000


Well bar is all well and good, except that it isn't importable for any of the software I use.

On the plus side, stuff I need to stick into tabular files rarely if ever has tabs in it, so TSV for me.
posted by adamv at 11:28 PM on May 20, 2000


I posted that Encyclopedia Brown link properly! whoohoo
posted by PWA_BadBoy at 11:59 PM on May 20, 2000


1. Baylink, thanks for that 32-year-old news flash! ;-)

2. Wren is lovely.

3. Comma-separated databases are only problematic if you're a loser stuck using regexp. ;-) If you're going to rant against a data format, rant against the bloated XML.
posted by dhartung at 2:52 AM on May 21, 2000


using BSV would make a lot of my (linux/bsd) shell scripts break, unless I went back and and "quoted 'inside' expressions". grep, sed, awk, perl, etc have control characters to recognize tabs, and most shells recognize CSV internally. I'm thinking there's an historical reason for the overwhelming popularity of CSV adn TSV.
posted by katchomko at 2:58 AM on May 21, 2000


I wonder if the 'historical reason' might have something to do with the 'DATA' or similar statements in some of the early computer languages. The DATA statement in many early BASICs allowed the programmer to include a section of comma-separated values inline within the program code, which were usually read into some kind of array by a loop. Other, mostly more recent, languages had much more sensible ways of initializing arrays and similar data structures.
posted by harmful at 8:10 AM on May 21, 2000


I've run into enough problems with data which contains commas that I've started using BSV without knowing anybody else was doing it. Of course, somebody who uses a lot of bars in their data is probably going to raise a cry against this technique. Personally, I'm looking forward to using XML, so I can use off-the-shelf parsers to handle more complex data.

The character that I really have a lot of problems with is the single quote (') as a string delimiter. With a fine Irish name like O'Keefe, I run into a lot of situations in which CGI scripts mishandle or outright choke on my name in a data entry form. I've also had no end of problems with my accursed work Lotus Notes address (Brennan.M.O'Keefe@domainname). By the way, I despise Lotus Notes in general; it may be good for something, but it's a lousy e-mail client.
posted by harmful at 8:17 AM on May 21, 2000


Indeed. Posting with NS6b1, lets see how this works...

Dan; don't thank me, thank Cam. ;-) (His Sullivan skin makes Moz almost livable, too, btw; gotta learn XML.

Why don't you like XML? Ok, stipulated, it's got some overhead, but overhead is *useful*. Bandwidth and diskspace are cheap; and I speak as someone who's had to parse crappy data formats for *many* years: I like XML.

I just wish people would learn that *all it is* is a way to wrap data...

posted by baylink at 11:13 AM on May 21, 2000


It's not that I dislike XML per se, it's just that as far as data formats go, it isn't traveling light by any means.
posted by dhartung at 1:59 PM on May 21, 2000


I think this is the dumbest argument I've ever heard. These are the kinds of arguments non-computer people imagine us having...
posted by premiumpolar at 5:29 PM on May 21, 2000


...sorry, I didn't mean to say "dumbest" it's not... I just think it's funny..
posted by premiumpolar at 5:30 PM on May 21, 2000


Of course, these are the arguments that we have so they don't have to.
posted by dhartung at 9:56 PM on May 21, 2000


« Older At the bottom of this article,   |   Congratulations Matt and Kay! Newer »


This thread has been archived and is closed to new comments