It's your letters page, so its not our fault....
Keen readers may have noticed that in last issue's editorial I explained that RISC World would be moving to 80min CDs from this issue onwards. This produced a number of letters of concern, including this from Michael Poole...
Hmmm. 80-min CDs from next time, eh? Well, if my second-hand 6x IBM drive won't read them, does that mean I've really got to get one of those 52x ones that this article says spins the discs up to 1500g at the edge and makes them fly to bits?
More to the point, there are zillions of CD drives in the shops here of all shapes, sizes and specifications, but it's going to be great fun trying to find out which ones do and don't operate with Acorn's software...
Well we have no choice. We simply cannot get sufficient stocks of the older type CDs any more. Having said that a 6x drive ought to be OK, it's the old type 2 x Cumana drives (and similar) that will fall over. Some of the original 4 x drives in very early RISC PC Computers may also have trouble. However readers should remember we are talking about drives that are at least eight years old, so they really should be due for replacement anyway. As for 52 speed CD drives destroying CDs, well it can happen if the CD is damaged, so if you look after CDs you have nothing to worry about.
Still, while we are on the subject of 52 speed drives....
I can't get into the latest CD through the !RWORLD application. The disc spins on and off and eventually comes up with a "can't find". Using Shift/Select on !RWORLD I get a gobbley gook directory content list in which directory sizes are listed and addresses, but no meaningful names and any attempt to shift enter gives an error message. I can get into the CD via the HTML directory and once in appear to be able to get into all the files that way. I appear to be able to read all the other directories OK. My CD drive was bought from ADPL and is a 52 speed one.
See what happens when you recommend a 52 speed APDL drive? Although in our defence I would like to say that most likely your CD is faulty for which we apologise. I can only suggest trying it in another machine to see if it works in that. If it doesn't, or another machine isn't available then return the CD to APDL for a replacement.
And while we are still talking CDs...
Here's the next batch of RISC World CD inlays (up to Oct 2002), in a zip archive.
Many thanks, we have put these in this issue's software directory. And now another unsolicited VirtualAcorn plug...
I no longer have an Acorn computer but do have a PC with VirtualA5000. My PC is a 2 year old Dell Inspiron. P3-600MHz with 100MHz main memory and 133MHz video memory. In terms of MIPS it is faster than a real A5000 but it is a huge deal faster when a lot of memory work is being done. (In general, using a lot of memory access it is faster than a RiscPC). Anyway, what I'm wondering is if you would be interested in little progs which would run best on VirtualA5000 running on a PC with at least my spec.
Bye for now
We are always delighted to receive submissions from readers. As many readers may remember I did try and revive the old Acorn User StarProgs column in RISC World, but regretfully we didn't receive enough submissions to make it worthwhile. Still if we do get submissions we will do our best to publish them. Oh, and just in case anyone thinks this letter was sent to VirtualAcorn rather than RISC World, it came in thought the StarProgs email address.
This is in response to your letters section where Eric Dobson wrote:
"With APDL developing and selling Ancestor+, and helping to produce
RiscWorld, why not include a series of articles on the subjects in
RiscWorld. Some may say that it could be interpreted as advertising for
APDL, but providing Hints & Tips for any RISC OS application should be
part of RiscWorld's task."
Let me tell a story to illustrate a problem:
My grandfather, who died some years before I was born so I cannot talk with him, left behind him some reams of notes on genealogy. All good stuff, you might think. But how do I know it is any good? He does not tell us where he got his information from, he gives us no idea which bits are solid and which are a tad risky (or RISCY?). So while it seems most impressive it is not a lot of use.
I have spent much of the last five or six years finding sources for his facts and, in the process, proved that some, most even, are valid and some not. And for every person I have found, I have documented where I found the information, so that anyone else who follows me will know how good my sources were and were to go to see what I actually found.
Genealogy must be reproducible, it must have within it the means to verify its content.
Regrettably Ancestor Plus does not include a file of source documents or books, so that one cannot easily cite the precise reference for the precise fact. On the other hand, all programs for other platforms do include facilities to create files of sources and then to conveniently enter references to these for every stone unturned. These other programs are professionally made for professionals. Ancestor Plus is not made or designed by a professional, or even a good amateur, genealogist - and shows it.
The RISC OS platform cries out for a good browser. Equally it cries out for a good genealogy program.
I would be delighted to assist anyone in drawing up specifications for a RISC OS based genealogy program. I am not a programmer, but I've done quite a bit of family history and put it on an internet site, see below (all prepared with RISCOS programs except for the main family tree).
First off I have to say that as one of the programming team that has worked on Ancestor+ I am a little biased. Ancestor+ does indeed allow you to attach a source file (with documentation etc) to each person in the family tree. You can also include photos, scanned documents or anything else that you need. This can be accessed using the little filing cabinet icon in the Edit person dialogue box. You can also use the blank fields in each persons details to store further information, for example links to other sources. We have included the latest Ancestor+ demo and the full HTML manual in this issue of RISC World. Would other readers care to comment?
And now a subject rather close to my heart....
I know that you are very familiar with the world of fonts. I am using Font
Fiend for a little job just at the moment. So I hope that you or the
readership of Risc World will be able to answer my question.
Is there any way of identifying an unfamiliar type face? I'm thinking about
true fonts, not the kind of thing that is probably the result of some designer
inventing his own lettering for the limited purposes of say an advertisement.
Sometimes in a magazine I will see the use of what is clearly a full font. I
will think "now that's a nice font, what's it's name and can I perhaps find a
copy somewhere?" I know that in the days before computers, when the number of
fonts held by a printer where limited, it was common to state at the beginning
of a book that say Garamond at 8pt was used, but even that has been
discontinued. In those days too a font could be identified very often by
picking out a distinctive letter, say a capital Q, and comparing it with fonts
that one knew, to obtain its name. I would have thought that computers were
ideal for this job of searching and identifying. Is there a program that can
Ah, this is one thorny issue. The main problem in identifying a typeface electronically is simple, pt size. One would think that at 8pt all fonts would be the same size, and you would be totally wrong. The typeface designer decides how big the font will be at 8pt, some fonts at 8pt, may be the same size as some other fonts at 6pt. Also depending on the pt size the letter shapes of the glyphs (characters) need to change to keep them legible. So a lower case "a" from garamond at 6pt, will not be the same shape as the same letter at 24pt. Not only that but the shape of the letter may well have been further altered on a computer by the font managers hinting system. The only reliable way to identify a font is with a typeface book. I have been working with fonts on and off for 13 years, so many I can identify by eye, for others I have to use a book.
I did at one time try to describe an algorithm to determine a fonts name. However the only way I could see of doing it was to store bitmaps of every font, on every platform, using every font rendering method and then compare the font being tested with each one. Even then I don't think it would work at all reliably.
I would like to to thank Richard Walker for his articles on networking. I would also like to ask him if he has any experience in setting up Linux to use a modem connected to a Internet/Printer Server box and if possible could he offer any help and advice.
I wonder if this is the same C Bass who used to teach at Leighton Park School in Reading? Still I will pass this on to Richard and of course forward any response back to you.
And now it seems that Michael Poole isn't the only reader who braves the editorial column in RISC World...
Where does one get a branded 21" monitor for £85 with guarantee?
Try Techno Supplies Limited (07748 682044). I also bought an 18" Taxan 975 TCO95 from them for 75ukp. They usually exhibit at small computer shows (both my monitors came from the Bracknell computer fair) and offer a two month guarantee on all the kit. So far both monitors have performed faultlessly, although both are now out of their two month periods so am I asking for it?
And now its the traditional RISC World letters page cock-up corner, where our authors tell me what part of last issue I made a dogs dinner of.
I just got the copy of Risc World forwarded on from the UK - thanks, nice to catch up with some other RISC OS news. I noticed you have credited the NTUsers article to Mike Choo. I wonder if this was intentional? I dont mind the pseudonym but I am in fact Mike Battersby and the mikechoo of the email address is a combination of mine and my wife's name. No problem except that any cheques need to be made out to my proper name to be paid in!
Sorry. I meant to correct this. I use e-mail addresses on the "flatplan" so I can easily hassle people who are late. Then by accident the article got credited using the wrong name. The articles in this issue are correctly credited and I will try and make sure it doesn't happen again.
I was somewhat worried, when re-reading my OS_ReadArgs article in Risc World for the first time more than a month after writing it, to find that I had apparently made some confusing and/or incorrect statements.... and somewhat taken aback, on checking my original, to then discover that any such confusion had been caused by punctuation marks that had 'gone missing' in transit. I didn't intend to state that keywords in the syntax string could be specified "in a predefined order which is a reasonable enough requirement" - it is not necessary for the order to be a reasonable requirement! - or that "if supplied [a string parameter] is the default element type", since it is the default element type whether it is supplied or not.
While it is undoubtedly the editorial privilege to amend my punctuation as desired, and while I'm aware that my natural style is more 'literary' than is the custom on Usenet and that I perhaps overuse semi-colons and dashes, I would point out that these punctuation marks are not in fact inserted by me at random but perform a fixed grammatical purpose. They serve to juxtapose two related but distinct sections of a sentence, and simply deleting them not only can, but frequently will, change its implications!
If they are to be removed, it is usually better to substitute them with a full stop and split the sentence back into its two constituent parts, retaining its intended meaning.
If I write any further articles I'll endeavour to keep my sentences shorter, thus reducing the scope for such misunderstanding.
I would also point out that there are *seven* errors on the Index page, including a misspelling of my own name - I suspect that, unlike the articles, this page doesn't get proof-read!
On an unrelated and perhaps more palatable subject: I can't get the image link insertion in the HTML_Link program working. The RunImage is supplied as an Absolute file, so I can't check the program for errors, but any attempt to insert the link in a file simply results in "Error 280 : Mistake". This sounds like a BASIC syntax error in the original source code rather than any error generated by the program due to bad input! The text link insertion works, though....
I was also surprised by David Holden's statement that in HTML, spaces before a closing tag are ignored. This doesn't seem to be the case for the four RISC OS browsers which I was able to test (Webite, Fresco, the Oregano demo and WebsterXL), so I wonder if it is a PC-specific phenomenon?
Right then working from the end of the letter back to the top. I think that Dave Holden may be incorrect about spaces before bold tags being ignored, certainly my experience so far has been that the are not ignored. The problem with the HTML_Link program is caused by "an undocumented keyboard input buffer stream error" (spelling mistake). A new version will be in this issues software directory.
I incorporated the space after closing tags because my RISC OS browser of choice, Browse, definitely does do this, and as all browsers which don't exhibit the problem follow the usual HTML convention of concatenating the two spaces into one there's usually no problem. Anyway, if you don't want to do it then don't add the space - it's not compulsary. DH.
As regards the editing I am sorry if I altered the meaning of the article. However I am afraid that your example "in a predefined order which is a reasonable enough requirement" isn't a good one as that is what you did say in the copy sent to me for RISC World. All I have done is replace a hyphen in your text with the words "which is". I do not see that this has altered the meaning in any way. I think the problem is that what you thought you had said and what you had actually said where not the same thing. I can only read what the author supplies, if the author thinks they have written one thing, but in fact they have written another then I am a bit stuck when it comes to editing. Also may I point out that if the article had been 'literary' then it wouldn't have needed editing.
I do my best to try and correct grammar and spelling when I can. Although as you have correctly noticed this didn't apply to the contents page (I will try harder in future).
That is it for this issue.