Wednesday, January 05, 2005

HE OPINES. HE WHINES! Check out Corey Pein's letter to Romenesko, which he's also posted on his own website. Pein's letter has already started to draw responses, and I suspect there will be quite a bit more tomorrow.

NEW IN THIS WEEK'S PHOENIX. Internet speculation about Bush's and Cheney's health poses a media dilemma. Also in this week's column: Mike Barnicle's Herald stint sours; what the sale of Slate means for online media; and Arthur Sulzberger Jr. changes his mind.

3 comments:

Anonymous said...

* Size 10EEE: wow. What is Dick, about 5'9,220? If I sliped on a pair of EEE's It would be like putting my foot in a shoebox. I hope the VP is ok.
* As for W's bulge, is it to much to expect to get an honest answer to an honest question? As Steve Martin would say "what the hell is that?" Like Bush's reason for the end of his flight career, I'd say they aren't going talk, which they are very good at.
* I can't remember when I've read Barnicle,or have cared to read him. But I'm glad to hear his multi-media career is chuging along.
* Does anybody remember what a big deal Lobel made over Bob Ryan's inadvertant "smack" remark over a ballplayer's wife? Lobel should stay away from lecturing anybody. Ryan who has really turned down the volume,made a mistake. Lobel has pathology issues.

Anonymous said...

To expand upon my previous comment after actually having RTFA's Corey Pein's piece ...

RaTHergate Redux - A Technical RetrospectiveCopyright © 2005 William D. Ricker All Rights ReservedCorey Pein's CJR deconstruction of the more rabid red bloggers and the MSM are on target, but repeats the error of the MSM in discarding the story with the messenger. Much of the MSM seems to have missed the point that regarding the purported "smoking CYA memos" this is a History of Technology investigation, not a he-said/she-said investigation. Only now that they're demonstrated likely forged does it becomes a he-said/she-said investigation - and it's past time for the "unamed" source (in Kinko's) to answer. It may well be that they were modern forged-copies to hide provenance (retypings, possibly in Kinko's?), or they may be total fantasy-forgeries.

As with Evolution, willful doubters (including some of Pein's critics) can pick at individual points and cite alternative views and claim "only a theory" -- but as such cultural-relativism is wrong for-Evolution doubters, it must be wrong for Font-doubters too.

60Min2 should have treated the memos as dubious until authenticated by both a chain of custody and document examination; they failed to do that. Based on the typographic evidence available, I must conclude the documents are highly dubious, at least for these copies.(At best they are modern retypings of actual memoranda, to hide an illicit chain of custody perhaps.) The evidentiary connections that Pein/CJR is missing is the combination of the typographic indicia; amongst the ravings of the rabid right there was some good investigative scholarship -- perhaps not journalism, but good technical history scholarship. One hopes facts matter in journalism, whether right or left.

Even if the original comment was on a raving righty blog site, and was phrased "for action", so what -- if the conclusion it jumped to was correct and checked out once it was checked out?

Aside for Bias disclosure & Qualifications -- I would love for the Rather story or the Globe story with the same facts but no documents to have "stuck" to Lt.Bush, even though it hadn't stuck 4 years prior either. But as an avocational printing historian (formerly webmaster and collections triage @ MuseumOfPrinting.org) and student of fonts, who has used computers for more years than most of these Bloggers and has carried an IBM Composing Selectric or two: once I saw the "smoking gun" memos at CBS, I knew the game was up on those documents at least.

Summary of the facts to date.Thought of in terms of a classical statistical experiment (for which we have too few test cases, but follow the terminology), the documents being producible in the era claimed is the null hypothesis, the norm that we wish to measure the likelihood of; if it is not compelling, that is the sign that something is wrong in the experiment. (Even if the null hypothesis checked out, a chain of custody would be nice too.)

There were some initial over-statements by the rabid righty bloggers. The rabid lefty counter-bloggers (and CBS/Media apologists) latched onto thes quickly. But we will see that they are actually harmless to the theory.

* that NO typewriter had a 'th' lig (ligature) then,

* that NO typewriter had proportional spacing, only top-end typesetting did;

* that NO typewriter had a Times-ish face then;

* that they can definitely detect/measure kerning and "MS Times Roman" in re-scanned faxes, regardless of the Nyquist criterion for scanning artifacts.

These are over-statements. We can check the historical facts about typewriters.

* ligs were available as factory options on many higher-end units, even then. Later, possibly then, some even had field-replaceable typebars. (E.g., my SCM portable from only a few years later had two such type-bar head snap-on replacement positions, one each left and right bars; no doubt st, nd, rd, th ligs was one of the several option-sets.

* and a few other (even rarer) typewriters (but not the Selectric) had a crude but effective form of proportional type that early.

* Although no typewriters had Times Roman, some typewriters had vaguely similar Roman-style faces.

* The reproduction of kerning is highly suggestive but at the resolution of rescanned faxes not conclusive; pixel-for pixel matching is impossible when it's been faxed and re-scanned, and was retracted by the fair and balanced blogger that made the claim.

However, none of these damages the MS Word hypothesis. If one actually cross-checks historical sources used to debunk these over-statements, one finds some relating facts that indicate that while any one of these nearly-true factoids can be false about a specific antique typewriter, no exemplar voids all the statements.

* The typewriters known to have 'th' lig options do not include the Proportional type typewriters and vice versa.

* Proportional typewriters had a less fluid proportional system (fewer larger units or character width) than in modern digital faces. Times Roman being copied from Monotype/Linotype faces does not use the full fluidity but would be still more fluid than IBM Executive or Composer. On rescanned faxes, under-sampled per Nyquist criterion, we can't measure closely, but we can see statistical emergent properties - e.g. line breaks. In fact, emergent properties of a line of type is how kerns and proportion were measured in typographic practice before we could open the font file and ask!

* While the IBM Composer aka IBM Selectric Composer or IBM Composing Selectric - which was not sold or normally used as a mere typewriter - could have produced a Times-ish proportional face, it did not kern, had very stiff proportions as above, and thus would likely miss at least one line break. Even the most Roman-styled Stock typeface didn't look quite Times-ish enough to match. It also didn't include TH in the standard Roman, according to IBM's Journal of R&D article Volume 12, Number 1, Page 26 (1968).Although they could have had a "special" typeball, of course. I wouldn't want to guess if it's too tall, but it feels too high for a "special", as well as too pale. (MSWord and other low-end products (compared to professional digital typesetting systems) have a known flaw among typographers from making psuedo-ligs from two letters in a smaller font which as a result have less "color".)

* The Margins are not to be ignored! Typographers know to examine the negative (or white) space. The relation between the default margin and the default font in MS WORD is the key to the reproduction match in LGF's experiment. The Margins and Font interact to produce line-breaks. A non-kerned font from a different family is statistically unlikely to achieve exactly the same word breaks as a kerned one.And this is with, LGF claimed - and you can test it at home with a vanilla normal.dot template - without adjusting the margins to fit. This is a Testable hypothesis: Change the font to Schoolbook or Courier and watch the line-breaks jump. Anyone who dismisses the experiment based on rejecting Pixel photoshopping is missing the point, examine the line-breaks. Do the experiment, or accept it.

The most interesting refutation of the IBM Selectric Composer or IBM Composing Selectric theory was also over-stated by someone who read a technical source but without jargon knowledge. On the composer, it was required to type each line twice only when not setting flush left rag right. However, you only bought a set of these if your publishing project required full justification and centering.

I delivered four of the IBM Composing Selectrics to the Museum Of Printing from the Belmont Press's old automated strike-on composition system. These had formerly been run by computer to set the body type for three community weeklies, with four similar keyboards providing text-to-tape, and an IBM computer handling justification.

The th, the 4, and most specifically the linebreaks are why the colonel didn't type these on his sergeant's newsletter composer.

-- Bill R

Anonymous said...

Dear Dan,

My longer *Technical* comments about Corey Pein's CJR deconstruction (longer than yesterday's quickie) are now posted on my website (until it moves?).

Cheers,

- Bill R