Å SigmaSheet does not return to the desktop neatly. If programs such as
Ovation are running, their Alt-key combinations will be lost when
SigmaSheet returns. To solve this, amend the SIGSHEET program as
follows. Look at the end and find DEFPROCexit and change it to read:
6.10
DEFPROCexit
6.10
*FX 4
6.10
ON:*FX 225,1
6.10
*RMReInit InternationalKeyboard
6.10
IF INKEYÖ2 : END : ELSE *Quit
6.10
John Waddell, Renfrewshire.ááA
Hints and Tips
6.9
Å Anti-aliased fonts Ö I find that anti-aliasing on a standard
resolution monitor gives a far too blurred image. My hint is to use the
outline version of the System font at an unexpanded point size. A bit of
experimentation is needed to get it right if you habitually use the
screen at some scale other than one to one but when you get it right,
the result is a nice fast, clean and clear font on any DTP or
wordprocessing package while you work on the text. Remember to change
the font to your preferred version for the final print. I am afraid I
canæt remember where I obtained the Outline System font.
6.9
On my Pentoms game, I changed the transfer coefficients to sharpen up
the edges of my fancy fonts but that does not work well/at all on
RISCáOS 3. Is there a way of getting a program to automatically check
which OS it is running under so that I can disable this facility where
it wonæt work? Or is there another way of sharpening RISCáOS 3 fonts?
Yes I know about turning off anti-aliasing but that is not quite the
same thing at all Ö bits of the letters are missed if you do that.á
Simon Anthony, Nottingham.
6.9
Å Backing up in one pass Ö The öNextò slot in the Task Manager can be
set from a !Boot file in RISC OS 3 to make backing up and copying of
full 800Kb discs occur in a single pass. Use the command
6.9
WimpSlot -next 864K
6.9
inside your !Boot file.á Barry Thompson,
6.9
Alternatively (for RISC OS 2 owners) you can allocate more memory to the
Next Task slot by issuing the Wimp_SlotSize SWI call from Basic. All
that you need is a Basic program containing the line:
6.9
SYS öWimp_SlotSizeò,-1,<slotsize>
6.9
where <slotsize> is in bytes (e.g. use 65536 for 64Kb.)
6.9
Simon Moy, NCS.
6.9
Å !Boot files are not strictly necessary within applications. They use
up disc space and slow down the opening of windows (especially over a
network). If an applicationæs !Boot file only contains the IconSprites
command, then this can be moved to the !Run file, and the !Boot file can
be deleted. The only effect is that the application will not show the
correct icon in the filer window until the application is actually run.á
Keith Harja, Derby.
6.9
Å Gods passwords Ö On this monthæs program disc is all you need for
getting the passwords for Gods. Andrew Ferguson, London N15.
6.9
Å Mono TIFF files Ö In reply to help for Ray Dawson in Archive 6.5 p18
for producing mono TIFF files from B&W scanned sprites. I have used
!MakeTiff version 2.00 by John Kortink to do exactly that. I examined
the resulting Tiff file using Graphics Workshop (under the emulator) and
Translator, and they both showed the converted sprite as a B&W image.á
Chris Bass, Grimsby.
6.9
Å Symbols for Impression Ö When using Impression, or even more so
Junior, it can sometimes be tedious to find a desired symbol from a font
such as Dingbats. Therefore, I have created a number of text files which
can be dragged into any Impression document. The format for such files
is ö« ò, without the quotes. This returns the type to its default size,
selects the desired font, inserts the character with ASCII code 174 and
resets the font to its default. N.B. The space at the end is essential.
This is so that one can do a <ctrl-right> to end of text. Without this
space, the font is not reset. I have given a number of pre-prepared ones
to Paul for inclusion on this monthæs program disc.á Neil Walker,
Norwich.
6.9
You need to be very careful with this one. I tried it on Impression 2.19
and it gave an internal error and Impression bombed out! This was
because I didnæt have the symbol font in my system. I tried it on
version 2.17 and it complained politely that it didnæt have Symbol font
and let me carry on. I presume that the difference is because
Impression, from 2.18 onwards, has no limit to the number of fonts it
can handle so that part of the code must have been changed. Donæt say we
didnæt warn you! Ed.
6.9
Å Testing printersæ on-line status Ö Here is a way of testing whether
the printer is on- or off-line:
6.9
10 *FX 3,10
6.9
20 PRINT CHR$(0);
6.9
30 a = INKEY(20)
6.9
40 a = ADVAL(-4)
6.9
50 *FX 3,0
6.9
60 *FX 21,3
6.9
70 IF a <> 1023 THEN
6.9
80 PRINT öPrinter not on lineò
6.9
90 ELSE
6.9
100 PRINT öPrinter on lineò
6.9
110 ENDIF
6.9
120 END
6.9
Paul Ribworth, Gloucester
6.9
Å Using high density discs Ö Until the advent of the A5000, all
Archimedes computers had an 800Kb 3╜ö disc drive so the problems of
different disc densities didnæt really arise. With the introduction of
the new A3000 and A4000 series, 1.6Mb drives are becoming more common
and some users are trying to save a few pence by formatting normal
density discs to 1.6Mb.
6.9
Users of BBC and Master computers learned that there is no point in
paying extra for 80 track discs as the cheaper 40 track were actually
identical. The same is NOT true of 800Kb and 1.6Mb discs. The magnetic
coating on high density discs uses finer particles than the coating used
on standard (double density) types. This is designed to accept higher
signal levels without distortion. When the higher capacity format is
used, the drive automatically uses a stronger magnetic field to imprint
the data on the discs surface.
6.9
Sometimes, good quality normal density discs can accept this level but
it does represent a severe overload. Even if all appears well at the
time, the larger particles used in the normal density coating are
frequently unable to retain the high level of magnetism required so it
can leak away in time making part of the data unreadable.
6.9
Also, you should not use high density discs in a normal density drive if
they have previously been formatted to 1.44 or 1.6Mb. The lower flux
levels used for the normal format may not be able to completely erase
the earlier data. This can then Ébreak throughæ and corrupt data saved
later at the lower flux levels used with the smaller format. It may
sometimes take more than a month for this to happen. So donæt be fooled
into thinking that everything is OK just because all seems to be well
when you first format the disc. Many Archimedes users who have
Éborrowedæ preformatted 1.44Mb discs from work to use in their 800Kb
drives have experienced this problem.
6.9
There is no reason why you cannot use a NEW high density disc at 800Kb,
it will just be more expensive.
6.9
Because of this, it is obviously useless to do what used to be common
practice with the old 40/80 track discs, namely try to format them to 80
tracks and if it didnæt work use them for 40 track. Once you have
attempted to format an 800Kb disc to 1.6 or 1.44 Mb there is no going
back. I have had some success using a tape head demagnetiser to
completely wipe out data on 800Kb discs which had been formatted to
1.6Mb so that they could be reformatted to 800Kb, but not everyone has
one of these.
6.9
PC drives all have a sensor which detects the extra hole in a high
density disc and wonæt allow you to specify the wrong format. For some
reason, Acorn didnæt implement this feature on the Archimedes although
the detector is present on all high density drives. It is therefore
possible, by accident, to format a normal density disc to 1.6Mb. On a
PC, to do this you need to drill a hole in the case of the disc so it
canæt happen unintentionally.
6.9
The most pernicious thing about this problem is that it can lead you
into a false sense of security. A normal density disc may format
perfectly to 1.6Mb and, for months, everything will appear to be well.
Then suddenly some of the data may become slightly corrupt, although the
disc may not actually appear to have any errors.
6.9
If your data is precious enough to be worth saving, itæs worth the extra
few pence that the proper discs cost. The only time that I have known
people use normal density discs successfully as high density was with
expensive branded discs costing about ú1.30 each. As this is
considerably more than the price of unbranded high density discs, it
seems a foolish risk to take.á David Holden, Sydenham.ááA
6.9
Hints and Tips
6.9
Å Anti-aliased fonts Ö I find that anti-aliasing on a standard resolution monitor gives a far too blurred image. My hint is to use the outline version of the System font at an unexpanded point size. A bit of experimentation is needed to get it right if you habitually use the screen at some scale other than one to one but when you get it right, the result is a nice fast, clean and clear font on any DTP or wordprocessing package while you work on the text. Remember to change the font to your preferred version for the final print. I am afraid I canæt remember where I obtained the Outline System font.
6.9
On my Pentoms game, I changed the transfer coefficients to sharpen up the edges of my fancy fonts but that does not work well/at all on RISCáOS 3. Is there a way of getting a program to automatically check which OS it is running under so that I can disable this facility where it wonæt work? Or is there another way of sharpening RISCáOS 3 fonts? Yes I know about turning off anti-aliasing but that is not quite the same thing at all Ö bits of the letters are missed if you do that.á Simon Anthony, Nottingham.
6.9
Å Backing up in one pass Ö The öNextò slot in the Task Manager can be set from a !Boot file in RISC OS 3 to make backing up and copying of full 800Kb discs occur in a single pass. Use the command
6.9
WimpSlot -next 864K
6.9
inside your !Boot file.á Barry Thompson,
6.9
Alternatively (for RISC OS 2 owners) you can allocate more memory to the Next Task slot by issuing the Wimp_SlotSize SWI call from Basic. All that you need is a Basic program containing the line:
6.9
SYS öWimp_SlotSizeò,-1,<slotsize>
6.9
where <slotsize> is in bytes (e.g. use 65536 for 64Kb.)
6.9
Simon Moy, NCS.
6.9
Å !Boot files are not strictly necessary within applications. They use up disc space and slow down the opening of windows (especially over a network). If an applicationæs !Boot file only contains the IconSprites command, then this can be moved to the !Run file, and the !Boot file can be deleted. The only effect is that the application will not show the correct icon in the filer window until the application is actually run.á Keith Harja, Derby.
6.9
Å Gods passwords Ö On this monthæs program disc is all you need for getting the passwords for Gods. Andrew Ferguson, London N15.
6.9
Å Mono TIFF files Ö In reply to help for Ray Dawson in Archive 6.5 p18 for producing mono TIFF files from B&W scanned sprites. I have used !MakeTiff version 2.00 by John Kortink to do exactly that. I examined the resulting Tiff file using Graphics Workshop (under the emulator) and Translator, and they both showed the converted sprite as a B&W image.á Chris Bass, Grimsby.
6.9
Å Symbols for Impression Ö When using Impression, or even more so Junior, it can sometimes be tedious to find a desired symbol from a font such as Dingbats. Therefore, I have created a number of text files which can be dragged into any Impression document. The format for such files is ö« ò, without the quotes. This returns the type to its default size, selects the desired font, inserts the character with ASCII code 174 and resets the font to its default. N.B. The space at the end is essential. This is so that one can do a <ctrl-right> to end of text. Without this space, the font is not reset. I have given a number of pre-prepared ones to Paul for inclusion on this monthæs program disc.á Neil Walker, Norwich.
6.9
You need to be very careful with this one. I tried it on Impression 2.19 and it gave an internal error and Impression bombed out! This was because I didnæt have the symbol font in my system. I tried it on version 2.17 and it complained politely that it didnæt have Symbol font and let me carry on. I presume that the difference is because Impression, from 2.18 onwards, has no limit to the number of fonts it can handle so that part of the code must have been changed. Donæt say we didnæt warn you! Ed.
6.9
Å Testing printersæ on-line status Ö Here is a way of testing whether the printer is on- or off-line:
6.9
10 *FX 3,10
6.9
20 PRINT CHR$(0);
6.9
30 a = INKEY(20)
6.9
40 a = ADVAL(-4)
6.9
50 *FX 3,0
6.9
60 *FX 21,3
6.9
70 IF a <> 1023 THEN
6.9
80 PRINT öPrinter not on lineò
6.9
90 ELSE
6.9
100 PRINT öPrinter on lineò
6.9
110 ENDIF
6.9
120 END
6.9
Paul Ribworth, Gloucester
6.9
Å Using high density discs Ö Until the advent of the A5000, all Archimedes computers had an 800Kb 3╜ö disc drive so the problems of different disc densities didnæt really arise. With the introduction of the new A3000 and A4000 series, 1.6Mb drives are becoming more common and some users are trying to save a few pence by formatting normal density discs to 1.6Mb.
6.9
Users of BBC and Master computers learned that there is no point in paying extra for 80 track discs as the cheaper 40 track were actually identical. The same is NOT true of 800Kb and 1.6Mb discs. The magnetic coating on high density discs uses finer particles than the coating used on standard (double density) types. This is designed to accept higher signal levels without distortion. When the higher capacity format is used, the drive automatically uses a stronger magnetic field to imprint the data on the discs surface.
6.9
Sometimes, good quality normal density discs can accept this level but it does represent a severe overload. Even if all appears well at the time, the larger particles used in the normal density coating are frequently unable to retain the high level of magnetism required so it can leak away in time making part of the data unreadable.
6.9
Also, you should not use high density discs in a normal density drive if they have previously been formatted to 1.44 or 1.6Mb. The lower flux levels used for the normal format may not be able to completely erase the earlier data. This can then Ébreak throughæ and corrupt data saved later at the lower flux levels used with the smaller format. It may sometimes take more than a month for this to happen. So donæt be fooled into thinking that everything is OK just because all seems to be well when you first format the disc. Many Archimedes users who have Éborrowedæ preformatted 1.44Mb discs from work to use in their 800Kb drives have experienced this problem.
6.9
There is no reason why you cannot use a NEW high density disc at 800Kb, it will just be more expensive.
6.9
Because of this, it is obviously useless to do what used to be common practice with the old 40/80 track discs, namely try to format them to 80 tracks and if it didnæt work use them for 40 track. Once you have attempted to format an 800Kb disc to 1.6 or 1.44 Mb there is no going back. I have had some success using a tape head demagnetiser to completely wipe out data on 800Kb discs which had been formatted to 1.6Mb so that they could be reformatted to 800Kb, but not everyone has one of these.
6.9
PC drives all have a sensor which detects the extra hole in a high density disc and wonæt allow you to specify the wrong format. For some reason, Acorn didnæt implement this feature on the Archimedes although the detector is present on all high density drives. It is therefore possible, by accident, to format a normal density disc to 1.6Mb. On a PC, to do this you need to drill a hole in the case of the disc so it canæt happen unintentionally.
6.9
The most pernicious thing about this problem is that it can lead you into a false sense of security. A normal density disc may format perfectly to 1.6Mb and, for months, everything will appear to be well. Then suddenly some of the data may become slightly corrupt, although the disc may not actually appear to have any errors.
6.9
If your data is precious enough to be worth saving, itæs worth the extra few pence that the proper discs cost. The only time that I have known people use normal density discs successfully as high density was with expensive branded discs costing about ú1.30 each. As this is considerably more than the price of unbranded high density discs, it seems a foolish risk to take.á David Holden, Sydenham.ááA
6.9
Using RISCáOSá3.10
6.9
Hugh Eagle
6.9
More encouragement for the faint-hearted!
6.9
Stuart Bell says: öHaving read the RISCáOS Column for the past few months, I approached the upgrade with caution, afraid of all sorts of installation and compatibility problems. I guess that by the nature of the things, itæs those people who encounter problems who write in, thereby giving a false impression of the true situation. So, if youære still with RISCáOS 2 and thinking about upgrading, be encouraged. Installing RISCáOS 3 on my A310, including removing the main board and installing the IFEL ROM carrier board took about an hour from start to finish, and I experienced no unforeseen problems.
6.9
öHaving been warned of problems with Compression 1.1, I de-compressed all files first. I await the free upgrade from CC. There is the reported problem with early versions of Font-FX and thatæs in hand.
6.9
öA couple of warnings, though: As well as powering up with <Delete> depressed, also make sure that youæve re-connected the keyboard first. (Donæt laugh ù it could have been you! That cost me several minutes of panic.) Secondly, alter the Font Size figure before loading your first Impression document, which otherwise can take several minutes to load with a 32Kb buffer.ò
6.9
Å Avie Electronics 4Mb A310 upgrade Ö The following information has come from Avie: LK12 should be altered to 1-3 and 2-4. Issue 1 PCBs do not have LK12 and require modification as follows:
6.9
Option 1: pins 2 and 24 on all of the ROMs are to be swapped. Remove pins 2 and 24 from each ROM socket by bending them up. Join pin 24 on all ROMs to GND. Join pin 2 on all ROMs to LA18 on IC28, pin 18.
6.9
Option 2: remove pin 2 from each stamped socket on PCB and strap a wire between the pins and LA18. Likewise, pin 24 can be strapped to GND after removing the pins in the stamped sockets.
6.9
Hardware problems
6.9
Watford Electronics ARM3 Ö When Colin Wood fitted a Watford Electronics ARM3 upgrade to his A410 it didnæt work. öThe screen flashed red/blue, but no beep. Just dead.ò As he had recently fitted RISCáOSá3, he decided to replace the RISCáOSá2 chips and everything seemed OK, but when he put the RISCáOSá3 chips back, again it didnæt work. Watfordæs help line couldnæt help, so he returned the ARM3 and since they were (again) out of stock he cancelled his order and got most (but not all, apparently) of his money back. Another ARM3, from CJE Microæs, worked first time.
6.9
Watford A310 RAM upgrades Ö Watford Electronics say that they are aware that their early A305/310 RAM upgrades do not work with RISCáOS 3 but they do not have any solution. They simply recommend that you buy one of their newer RAM upgrades instead because they do work with RISCáOS 3. Fortunately, IFEL have stepped into the breach. They can offer new PAL chips to plug into the Watford boards that will solve the problem. These PAL chips are ú3 +VAT each from IFEL and you need to state whether you have the 2Mb or the 4Mb version. (This refers to the upgrades that consisted of two PCBs with a ribbon cable along the RHS of the machine.)
6.9
Program problems
6.9
QuicKey clash Ö J. Nichols reports that, if QuicKey is loaded from within a !Boot file, it can remove the *Opt 4,2 status from the booted disc resulting in a disc which no longer runs a !Boot file.
6.9
!Edit ù Following last monthæs reports of a öbugò in Edit we now have the following from Tom Rank: öThere seemed to be a rare problem with Edit the other day; my text (exported from Ovation) was displayed in a most corrupted way, with a large amount of blank screen to start with and then disintegrating words. I thought it was a problem with Ovation. However, when I tried to show it to John Wallace of Risc Developments at BETT, the file displayed without problems, as it did back on my own machine the next day. Any idea why?ò
6.9
Fun School 4 (for the under 5s) ù Tom Rank found that this ran under RISCáOSá3.00 but not 3.10, but Europress Software supplied an updated disc very promptly which now works.
6.9
Freddy Teddy ù Topologika supplied a new version to Tom Rank on the spot at BETT!
6.9
Ovation ù Tom Rank says you are urged to obtain 1.37S (dated 17 January 1993). This overcomes the saving with <F3> problem ù which really was a problem, so do upgrade. They have also restored the Éselect allæ shortcut Ö but only if you use the right-hand <Ctrl> key along with A (to avoid the problem of catching <Ctrl-A> instead of <Ctrl-S> or just <A>).
6.9
This version also contains three new printing options, not documented: Fit, Centre and Registration marks. These should be self explanatory, but Tom is not quite sure about them!
6.9
Break147 and SuperPool ù D. S. Allen got these to work using the information provided by David Holden in the April column. He writes: öIn Break147 there is one Basic program called !Snook which contains the line
6.9
90SYSá26,64,64.
6.9
A REM statement in front of the SYS will cure the problem and the program will then run under RISCáOSá3.1.
6.9
öSimilarly, in SuperPool there is a Basic program !Pool in the Data directory in which line 60 has the same SYSá26,64,64 statement.
6.9
öNote that to make these changes, people will require the read/write version of Mark Smithæs excellent ArcFS program.ò (Or, presumably, an equivalent utility such as SparkFS.)
6.9
Superior Golf / Construction Set ù We reported in Archive 6.7 p43 that Dave Wilcox had found that these did not work in RISCáOSá3.10. Rob Brown, however, says he has had no problems with them.
6.9
Corruption ù Rob Brown says that this can be made to work by modifying its !Run file to kill the RISCáOSá3 Shared C Library module and load the very old version of the module supplied with the game. To do this, include at the beginning of the !Run file the line:
6.9
RMKill SharedCLibrary
6.9
and then add at the end of the file the following two lines:
6.9
RMKill SharedCLibrary
6.9
RMLoad SharedCLibrary
6.9
to remove the old version and reload the RISCáOSá3 version of the module. He points out that it may well be advisable to Quit any other applications that may be using the RISCáOSá3 Shared C Library module before running the game.
6.9
Freddyæs Folly ù Rob Brown advises that this game contains a !Run file (in Basic) which includes an automatic Éauto-configurationæ of the computeræs memory as well as setting up a number of initial program parameters. The various auto-configuration elements should be deleted from this file. A further complication is that the disc is copy protected, but it should be possible to rename the original !Run file (to say !RunOLD) and re-save a new !Run file containing only the initial program parameters. He also advises that the game assumes that the default disc drive is 0. For users with a hard disc this can easily be set by typing:
6.9
Drive 0
6.9
at the command line (accessed by pressing function key <F12>) before starting the game. Alternatively, the gameæs !Run file can be appropriately amended by including the ö<Obey$Dir>ò command.
6.9
NOTE: the advice on Corruption and Freddyæs Folly is included in the program compatibility chart which Rob Brown has compiled and which has been included on the Archive monthly program disc. He will welcome any information which will help him to make the chart as comprehensive and up-to-date as possible. His address is öValtaneeò, Brighton Road, Lower Kingswood, Tadworth, Surrey, KT20 6UP.
6.9
Aldebaran ù (see Archive 6.8 p23) Atle Mjelde Bσrdholt has contacted the author and has learned that, on his A310 with IFELæs 2Mb upgrade, he has to add 8Kb to the WimpSlot values in the !Aldebaran.!Run file (from 624Kb to 632Kb). Then everything works fine. He says this might be due to the RAM upgrade: Aldebaran seems to work fine on any other Archimedes with RISCáOSá3.1. He adds a final bit of advice: öGet a copy!ò
6.9
Printing
6.9
Printing to file
6.9
Tom Rank has found that the hint from Simon Moy in Archive 6.2.7 about editing the number of text lines seems redundant under RISCáOSá3.10, and that printing to a file (at least with the Epson LQ-860 driver) resulted in a printout which aborted with an öaddress exception at &038B5BE0ò, a frozen printer and an incomplete document (the last few millimetres were lost). This is particularly irritating as printing to a file is the only way, apart from buying a direct laser printer or a CC Turbo Driver, to avoid the long wait for print out. (How about using the PROdriver shell for use with Acorn drivers? See the review on page 43. Ed.)
6.9
Using fanfold paper
6.9
Tom Rank comments: öFeeding single sheets into a printer soon becomes a pain, but sheetfeeders come expensive, so A4 fanfold paper seemed a good idea (even though itæs about four or five times more expensive than better quality photocopy paper at a discount stationer!). However, with RISCáOS 3.00, I could never get the settings right, so the print always crept up or down the page until it crossed the fold. With RISCáOS 3.10, Iæve finally got it to work, whether by fluke Iæm not sure. Using the LQ-860 printer driver with my Epson LQ-550 printer, the correct settings from the paper sizes menu are: top margin 8mm and bottom margin 15mm; save these with a name like ÉA4 Fanfoldæ and the margins can be automatically set each time. These settings seem to work OK with single sheets too.ò
6.9
Filing Systems
6.9
Beebug 5╝ö disc interface
6.9
Stuart Bell writes: öI got my Beebug 5╝ö disc interface (date 1989) to work by setting the step rate for an 80 track double-sided drive to 12ms (lower values may well work; Iæm just cautious), and whereas RISCáOSá2 worked with öReadyò (switch 6 on the interface) supplied by the on-board PAL, RISCáOSá3 works more reliably with it passed through from the drive.ò
6.9
Dismounting between filing systems
6.9
According to Mike Sawle, applications which frequently switch between filing systems seem to require the *DISMOUNT command before selecting an alternative filing system in order to ensure correct execution. Common error messages are Ambiguous Disc Name and Bad Error Block.
6.9
Miscellaneous Hints & Tips
6.9
Å Faster copying, etc. ù Rob Davison recommends that if there is nothing else you want to do when copying files, then click <Menu> over the filer action window and choose ÉFasteræ.
6.9
Å In the midnight hour ù Atle Mjelde Bσrdholt found that, whenever he switched his computer on between midnight and 1áa.m., a day was added to the internal clock. He then found that this seemed to be caused by the presence in his !Boot file of the command *SetáSys$Yeará1993 (which he had included because he has some programs that set the year back to 1988). When this command is removed, the midnight hour problem goes away. Does anyone know the reason?
6.9
Å Fan filters and the A5000 ù Tom Rank found that, although magazines mention the need to change the filter from time to time, there was no mention of how to do this and on fitting the RISCáOS 3.10 upgrade he could see no filter anyway. Is there one, he asks? The Acorn representative at BETT told him that the design of the machine made it superfluous; he couldnæt quite follow why, but it was something to do with the direction of airflow.
6.9
Å Flippinæ objects in Draw ù Having travelled all the way from Derbyshire to the Acorn User Show, Tom Rank öasked the man on one of the other magazinesæ stall if there was an equivalent in Draw to the Éflipæ option in Paint, as itæs sometimes useful to turn a picture round to face the other way. The expert, although helpful about other items, couldnæt tell me, but my son revealed the following tip, which he said he thought was obvious. Must talk to him more often....
6.9
öTo flip an object in Draw, select it and, from the Étransformæ menu (RISCáOS 3), set x scale (or y scale or both) to Ö1 (minus 1). This also works in the RISCáOS 2 version of Draw but there the Étransformæ option is chosen from the Select menu.ò
6.9
Å Lost lines in Edit Ö Eric Ayers has found that double-clicking on large text files to load them into Edit has caused several lines to be missing from the middle of the text. Apparently, this does not happen if he drags the files to the Edit icon on the icon bar. Any ideas?
6.9
Å More fonts in RISC-OS 3 Ö Gary Atkinson of KudlianSoft recommends the following additions to the !Fonts.!Boot and !Fonts.!Run files to enable one or more separate font directories to be added to and removed from the font path simply by double-clicking on the fonts application directory.
6.9
The !Fonts.!Boot file should be amended as follows:
6.9
| Boot file for !Fonts.
6.9
| Nothing happens if you put this in | a RISCáOS 2 machine except that
6.9
| the !Fonts sprite is set up Ö this | means that, when a RISCáOS2 !Fonts
6.9
| directory is seen the !Boot file
6.9
| will not be run; double clicking
6.9
| on the appropriate !Fonts will
6.9
| force installation.
6.9
|
6.9
| Take care when merging RISCáOS 2
6.9
| fonts with these (RISCáOS 3) fonts
6.9
| - many fonts are contained in the
6.9
| RISCáOS 3 ROM and thus do not
6.9
| appear in !Fonts; normally you
6.9
| should not copy the corresponding
6.9
| RISCáOS 2 fonts!
6.9
Iconsprites <Obey$Dir>.!Sprites
6.9
RMEnsure FontManager 2.89 Obey
6.9
| The next command boots up your
6.9
| normal font directory
6.9
FontInstall <Obey$Dir>.
6.9
SetEval MoreFonts 0
6.9
The !Fonts.!Run file should be amended as follows:
6.9
| !Run file for !Fonts,
6.9
| version 1.05 (20-Aug-91)
6.9
RMEnsure FontManager 2.89 Error These fonts are for use with
6.9
RISCáOS 2.12 or later
6.9
|Obey <Obey$Dir>.!Boot
6.9
| The extra *FontInstall means that,
6.9
| if you double-click the Fonts
6.9
| directory, all such directories
6.9
| will be re-searched for any new
6.9
| fonts.
6.9
SetEval MoreFonts MoreFonts EOR 1
6.9
IF MoreFonts=1 THEN FontInstall <Obey$Dir>. ELSE FontRemove <Obey$Dir>.
6.9
|FontInstall
6.9
The !Boot file is only run on the first time that the filer sees the !Fonts application, hence the initialisation of MoreFonts to zero. Every time the directory is double-clicked, the !Run file will toggle the value of this variable between 0 and 1. This is done in the line:
6.9
SetEval MoreFonts MoreFonts EOR 1
6.9
If the value is 1, the FontInstall command adds the new directory to the font path and initialises all of the fonts. If the value is 0, the FontRemove command removes the temporary font directory from the font path.
6.9
If you require more than one spare directory, you will need to use different variables, such as MoreFonts2 or FancyFonts.
6.9
In practice, it is better to have a directory called fonts inside your !Fonts application which contains your preferred normal fonts at switch-on, and a directory called spare (for example) which contains the set of fonts that you wish to switch in and out. Then in your !Boot file, the FontInstall command becomes
6.9
FontInstall <Obey$Dir>.fonts.
6.9
and the IF...THEN statement in the !Run file has <Obey$Dir>.spare. in place of the two occurrences of <Obey$Dir>. .
6.9
Miscellaneous Comments
6.9
FontEd
6.9
When Tom Rank spoke to an Acorn representative he was told that they were not contemplating the release of FontEd and that it is NOT public domain. An early version got there by mistake via SID and it really ought not, he said, to be distributed by PD libraries.
6.9
The Acorn representative acknowledged that Acorn were unlikely to do anything about this situation, as it was öonly an early version and wonæt work on RISCáOS 3ò. According to Tom this is partially true; a friend of his has had problems with version 0.28, but at least some of the time it runs on an A5000. However, he observes that it is a pity that Acorn actually seem to be discouraging anyone else from working on fonts. (There is a version of FontEd on Careware 7, released in May 1990, but it has been updated since then. Acorn havenæt objected, presumably because of the charitable aspect. Ed.)
6.9
Screen blanking during floppy disc access
6.9
Stuart Bell writes: öMany things are very much more elegant in RISCáOS 3 but with one notable exception. I couldnæt believe my eyes when I saw what happens to the screen during floppy disc accessing. Whether my ARM3 cache is dis- or enabled, in mode 24 the screen either flashes or is blanked when loading, saving or formatting. If RISCáOS 2 didnæt need to do this, why is it allowed now? (It did, in fact, happen in mode 21 under RISCáOSá2 ù HE.) Some re-coding is needed, even if the resultant code is a little slower. More significantly, who was the Acorn person in authority who thought that it is acceptable in 1993 to produce a computer which professes to be a Éworld-beateræ which, when using floppy discs, looks like the Sinclair Spectrum (or was it the ZX81?) reading from cassette tape circa 1983? This just looks so tacky. I know that serious use demands a hard disc, but first impressions count for a lot, and a lot of demonstrations will be on floppy-based A3010æs. This is not good enough, Acorn.ò
6.9
Basic program compaction errors
6.9
R. W. Darlington writes: öI am writing on a matter of utmost urgency.
6.9
öThere seems to be a growing urge amongst programmers to use a program compactor on Basic RunImage files. Unfortunately, I know of no Basic compactor program that does not exhibit a certain disastrous fault.
6.9
öIf, for instance, a Basic program contains any floating point or pseudo variables followed by a Basic keyword or statement, then when it is compacted and the space between the variable and the keyword is removed, that line will generate an error.
6.9
öFor instance, if the program contained the line
6.9
10 IF TRUE THEN VDU7
6.9
which when compacted becomes
6.9
10IFTRUETHENVDU7
6.9
the lack of space between TRUE and THEN will generate an error.
6.9
öThe same is true for any line containing a floating point variable:
6.9
10 ha=TRUE
6.9
20 IF ha THEN VDU7
6.9
will crash when compacted because of a lack of space between ha and THEN.
6.9
öI have come across numerous examples of RunImage files which have these errors in them. One of them is Acornæs new RISCáOS 3 !SciCalc. Others are new versions of Translator and Creator.
6.9
öPlease can you draw Archive readersæ attention to these problems as soon as possible? Or, if any reader knows of a Basic compactor that does not produce these errors, please can they let us know where it may be obtained?ò
6.9
Matters Arising
6.9
Setting Caps Lock (*FX 202)
6.9
Unfortunately, the advice I gave last month about using *FX 202 may have been a bit confusing because, after explaining that Peter Young wanted a command to turn Caps Lock on, I then described how to turn it off! The command to turn Caps Lock on, leaving the other settings unchanged, is *FXá202,0,239.
6.9
The shortened form *FXá202,32 should turn Caps Lock on, but may have side effects.
6.9
Note also that, to make sure that the LED indicators are correctly updated after a *FXá202 command, it is a good idea to follow it with the simple command *FXá118.
6.9
The bits of the keyboard status byte have the following meanings:
6.9
bit value meaning
6.9
0 (1) 0 Alt off
6.9
1 Alt on
6.9
1 (2) 0 Scroll Lock off
6.9
1 Scroll Lock on
6.9
2 (4) 0 Num Lock on
6.9
1 Num Lock off
6.9
3 (8) 0 Shift off
6.9
1 Shift on
6.9
4 (16) 0 Caps Lock on
6.9
1 Caps Lock off
6.9
5 (32) (öNormally setò)
6.9
6 (64) 0 Ctrl off
6.9
1 Ctrl on
6.9
7 (128) 0 Shift enable off
6.9
1 Shift enable on
6.9
(This information is taken from the RISC¡OSá2 Programmersæ Reference Manual.)
6.9
Boot-up options (Archive 6.8 p26)
6.9
Peter Young has suggested that it might be better to use INKEY$ with a delay of, say, 5 to 10 seconds instead of GET$ so that, if nobody presses a key, the computer will continue with the boot file after a short delay.
6.9
Unexpected messages from IDE filing system (Archive 6.8 p24)
6.9
Peter Young (who has a Beebug IDE interface) has confirmed that, even after creating space for a directory cache, he still gets unexpected messages unless he uses full pathnames. For instance, if he starts with the commands
6.9
*DIR IDEFS::IDEDisc4.$.BASIC.B
6.9
*CAT
6.9
he gets a catalogue of the subdirectory ÉBæ as expected. However, if he uses the abbreviated pathname instead:
6.9
*DIR $.BASIC.B
6.9
*CAT
6.9
he gets a catalogue of the directory ÉBASICæ even though the heading of the directory listing correctly says that the CSD is IDEFS::IDEDisc4.$. BASIC.B. The computeræs confusion is further demonstrated by the fact that if he then enters:
6.9
*DIR ^
6.9
*CAT
6.9
it replies Directory É@æ not found.
6.9
ÉInterruptæ module (Archive 6.6 p55)
6.9
In March, we reported M. P. Sawleæs request for help in getting the Interrupt module (Archive Shareware disc 6) working under RISCáOSá3, but since then neither he nor we have heard anything. Does no-one else use this module? Can its author help?
6.9
Using RISCáOSá3.10
6.9
Hugh Eagle
6.9
More encouragement for the faint-hearted!
6.9
Stuart Bell says: öHaving read the RISCáOS Column for the past few
months, I approached the upgrade with caution, afraid of all sorts of
installation and compatibility problems. I guess that by the nature of
the things, itæs those people who encounter problems who write in,
thereby giving a false impression of the true situation. So, if youære
still with RISCáOS 2 and thinking about upgrading, be encouraged.
Installing RISCáOS 3 on my A310, including removing the main board and
installing the IFEL ROM carrier board took about an hour from start to
finish, and I experienced no unforeseen problems.
6.9
öHaving been warned of problems with Compression 1.1, I de-compressed
all files first. I await the free upgrade from CC. There is the reported
problem with early versions of Font-FX and thatæs in hand.
6.9
öA couple of warnings, though: As well as powering up with <Delete>
depressed, also make sure that youæve re-connected the keyboard first.
(Donæt laugh ù it could have been you! That cost me several minutes of
panic.) Secondly, alter the Font Size figure before loading your first
Impression document, which otherwise can take several minutes to load
with a 32Kb buffer.ò
6.9
Å Avie Electronics 4Mb A310 upgrade Ö The following information has come
from Avie: LK12 should be altered to 1-3 and 2-4. Issue 1 PCBs do not
have LK12 and require modification as follows:
6.9
Option 1: pins 2 and 24 on all of the ROMs are to be swapped. Remove
pins 2 and 24 from each ROM socket by bending them up. Join pin 24 on
all ROMs to GND. Join pin 2 on all ROMs to LA18 on IC28, pin 18.
6.9
Option 2: remove pin 2 from each stamped socket on PCB and strap a wire
between the pins and LA18. Likewise, pin 24 can be strapped to GND after
removing the pins in the stamped sockets.
6.9
Hardware problems
6.9
Watford Electronics ARM3 Ö When Colin Wood fitted a Watford Electronics
ARM3 upgrade to his A410 it didnæt work. öThe screen flashed red/blue,
but no beep. Just dead.ò As he had recently fitted RISCáOSá3, he decided
to replace the RISCáOSá2 chips and everything seemed OK, but when he put
the RISCáOSá3 chips back, again it didnæt work. Watfordæs help line
couldnæt help, so he returned the ARM3 and since they were (again) out
of stock he cancelled his order and got most (but not all, apparently)
of his money back. Another ARM3, from CJE Microæs, worked first time.
6.9
Watford A310 RAM upgrades Ö Watford Electronics say that they are aware
that their early A305/310 RAM upgrades do not work with RISCáOS 3 but
they do not have any solution. They simply recommend that you buy one of
their newer RAM upgrades instead because they do work with RISCáOS 3.
Fortunately, IFEL have stepped into the breach. They can offer new PAL
chips to plug into the Watford boards that will solve the problem. These
PAL chips are ú3 +VAT each from IFEL and you need to state whether you
have the 2Mb or the 4Mb version. (This refers to the upgrades that
consisted of two PCBs with a ribbon cable along the RHS of the machine.)
6.9
Program problems
6.9
QuicKey clash Ö J. Nichols reports that, if QuicKey is loaded from
within a !Boot file, it can remove the *Opt 4,2 status from the booted
disc resulting in a disc which no longer runs a !Boot file.
6.9
!Edit ù Following last monthæs reports of a öbugò in Edit we now have
the following from Tom Rank: öThere seemed to be a rare problem with
Edit the other day; my text (exported from Ovation) was displayed in a
most corrupted way, with a large amount of blank screen to start with
and then disintegrating words. I thought it was a problem with Ovation.
However, when I tried to show it to John Wallace of Risc Developments at
BETT, the file displayed without problems, as it did back on my own
machine the next day. Any idea why?ò
6.9
Fun School 4 (for the under 5s) ù Tom Rank found that this ran under
RISCáOSá3.00 but not 3.10, but Europress Software supplied an updated
disc very promptly which now works.
6.9
Freddy Teddy ù Topologika supplied a new version to Tom Rank on the spot
at BETT!
6.9
Ovation ù Tom Rank says you are urged to obtain 1.37S (dated 17 January
1993). This overcomes the saving with <F3> problem ù which really was a
problem, so do upgrade. They have also restored the Éselect allæ
shortcut Ö but only if you use the right-hand <Ctrl> key along with A
(to avoid the problem of catching <Ctrl-A> instead of <Ctrl-S> or just
<A>).
6.9
This version also contains three new printing options, not documented:
Fit, Centre and Registration marks. These should be self explanatory,
but Tom is not quite sure about them!
6.9
Break147 and SuperPool ù D. S. Allen got these to work using the
information provided by David Holden in the April column. He writes: öIn
Break147 there is one Basic program called !Snook which contains the
line
6.9
90SYSá26,64,64.
6.9
A REM statement in front of the SYS will cure the problem and the
program will then run under RISCáOSá3.1.
6.9
öSimilarly, in SuperPool there is a Basic program !Pool in the Data
directory in which line 60 has the same SYSá26,64,64 statement.
6.9
öNote that to make these changes, people will require the read/write
version of Mark Smithæs excellent ArcFS program.ò (Or, presumably, an
equivalent utility such as SparkFS.)
6.9
Superior Golf / Construction Set ù We reported in Archive 6.7 p43 that
Dave Wilcox had found that these did not work in RISCáOSá3.10. Rob
Brown, however, says he has had no problems with them.
6.9
Corruption ù Rob Brown says that this can be made to work by modifying
its !Run file to kill the RISCáOSá3 Shared C Library module and load the
very old version of the module supplied with the game. To do this,
include at the beginning of the !Run file the line:
6.9
RMKill SharedCLibrary
6.9
and then add at the end of the file the following two lines:
6.9
RMKill SharedCLibrary
6.9
RMLoad SharedCLibrary
6.9
to remove the old version and reload the RISCáOSá3 version of the
module. He points out that it may well be advisable to Quit any other
applications that may be using the RISCáOSá3 Shared C Library module
before running the game.
6.9
Freddyæs Folly ù Rob Brown advises that this game contains a !Run file
(in Basic) which includes an automatic Éauto-configurationæ of the
computeræs memory as well as setting up a number of initial program
parameters. The various auto-configuration elements should be deleted
from this file. A further complication is that the disc is copy
protected, but it should be possible to rename the original !Run file
(to say !RunOLD) and re-save a new !Run file containing only the initial
program parameters. He also advises that the game assumes that the
default disc drive is 0. For users with a hard disc this can easily be
set by typing:
6.9
Drive 0
6.9
at the command line (accessed by pressing function key <F12>) before
starting the game. Alternatively, the gameæs !Run file can be
appropriately amended by including the ö<Obey$Dir>ò command.
6.9
NOTE: the advice on Corruption and Freddyæs Folly is included in the
program compatibility chart which Rob Brown has compiled and which has
been included on the Archive monthly program disc. He will welcome any
information which will help him to make the chart as comprehensive and
up-to-date as possible. His address is öValtaneeò, Brighton Road, Lower
Kingswood, Tadworth, Surrey, KT20 6UP.
6.9
Aldebaran ù (see Archive 6.8 p23) Atle Mjelde Bσrdholt has contacted the
author and has learned that, on his A310 with IFELæs 2Mb upgrade, he has
to add 8Kb to the WimpSlot values in the !Aldebaran.!Run file (from
624Kb to 632Kb). Then everything works fine. He says this might be due
to the RAM upgrade: Aldebaran seems to work fine on any other Archimedes
with RISCáOSá3.1. He adds a final bit of advice: öGet a copy!ò
6.9
Printing
6.9
Printing to file
6.9
Tom Rank has found that the hint from Simon Moy in Archive 6.2.7 about
editing the number of text lines seems redundant under RISCáOSá3.10, and
that printing to a file (at least with the Epson LQ-860 driver) resulted
in a printout which aborted with an öaddress exception at &038B5BE0ò, a
frozen printer and an incomplete document (the last few millimetres were
lost). This is particularly irritating as printing to a file is the only
way, apart from buying a direct laser printer or a CC Turbo Driver, to
avoid the long wait for print out. (How about using the PROdriver shell
for use with Acorn drivers? See the review on page 43. Ed.)
6.9
Using fanfold paper
6.9
Tom Rank comments: öFeeding single sheets into a printer soon becomes a
pain, but sheetfeeders come expensive, so A4 fanfold paper seemed a good
idea (even though itæs about four or five times more expensive than
better quality photocopy paper at a discount stationer!). However, with
RISCáOS 3.00, I could never get the settings right, so the print always
crept up or down the page until it crossed the fold. With RISCáOS 3.10,
Iæve finally got it to work, whether by fluke Iæm not sure. Using the
LQ-860 printer driver with my Epson LQ-550 printer, the correct settings
from the paper sizes menu are: top margin 8mm and bottom margin 15mm;
save these with a name like ÉA4 Fanfoldæ and the margins can be
automatically set each time. These settings seem to work OK with single
sheets too.ò
6.9
Filing Systems
6.9
Beebug 5╝ö disc interface
6.9
Stuart Bell writes: öI got my Beebug 5╝ö disc interface (date 1989) to
work by setting the step rate for an 80 track double-sided drive to 12ms
(lower values may well work; Iæm just cautious), and whereas RISCáOSá2
worked with öReadyò (switch 6 on the interface) supplied by the on-board
PAL, RISCáOSá3 works more reliably with it passed through from the
drive.ò
6.9
Dismounting between filing systems
6.9
According to Mike Sawle, applications which frequently switch between
filing systems seem to require the *DISMOUNT command before selecting an
alternative filing system in order to ensure correct execution. Common
error messages are Ambiguous Disc Name and Bad Error Block.
6.9
Miscellaneous Hints & Tips
6.9
Å Faster copying, etc. ù Rob Davison recommends that if there is nothing
else you want to do when copying files, then click <Menu> over the filer
action window and choose ÉFasteræ.
6.9
Å In the midnight hour ù Atle Mjelde Bσrdholt found that, whenever he
switched his computer on between midnight and 1áa.m., a day was added to
the internal clock. He then found that this seemed to be caused by the
presence in his !Boot file of the command *SetáSys$Yeará1993 (which he
had included because he has some programs that set the year back to
1988). When this command is removed, the midnight hour problem goes
away. Does anyone know the reason?
6.9
Å Fan filters and the A5000 ù Tom Rank found that, although magazines
mention the need to change the filter from time to time, there was no
mention of how to do this and on fitting the RISCáOS 3.10 upgrade he
could see no filter anyway. Is there one, he asks? The Acorn
representative at BETT told him that the design of the machine made it
superfluous; he couldnæt quite follow why, but it was something to do
with the direction of airflow.
6.9
Å Flippinæ objects in Draw ù Having travelled all the way from
Derbyshire to the Acorn User Show, Tom Rank öasked the man on one of the
other magazinesæ stall if there was an equivalent in Draw to the Éflipæ
option in Paint, as itæs sometimes useful to turn a picture round to
face the other way. The expert, although helpful about other items,
couldnæt tell me, but my son revealed the following tip, which he said
he thought was obvious. Must talk to him more often....
6.9
öTo flip an object in Draw, select it and, from the Étransformæ menu
(RISCáOS 3), set x scale (or y scale or both) to Ö1 (minus 1). This also
works in the RISCáOS 2 version of Draw but there the Étransformæ option
is chosen from the Select menu.ò
6.9
Å Lost lines in Edit Ö Eric Ayers has found that double-clicking on
large text files to load them into Edit has caused several lines to be
missing from the middle of the text. Apparently, this does not happen if
he drags the files to the Edit icon on the icon bar. Any ideas?
6.9
Å More fonts in RISC-OS 3 Ö Gary Atkinson of KudlianSoft recommends the
following additions to the !Fonts.!Boot and !Fonts.!Run files to enable
one or more separate font directories to be added to and removed from
the font path simply by double-clicking on the fonts application
directory.
6.9
The !Fonts.!Boot file should be amended as follows:
6.9
| Boot file for !Fonts.
6.9
| Nothing happens if you put this in | a RISCáOS 2 machine except that
6.9
| the !Fonts sprite is set up Ö this | means that, when a RISCáOS2
!Fonts
6.9
| directory is seen the !Boot file
6.9
| will not be run; double clicking
6.9
| on the appropriate !Fonts will
6.9
| force installation.
6.9
|
6.9
| Take care when merging RISCáOS 2
6.9
| fonts with these (RISCáOS 3) fonts
6.9
| - many fonts are contained in the
6.9
| RISCáOS 3 ROM and thus do not
6.9
| appear in !Fonts; normally you
6.9
| should not copy the corresponding
6.9
| RISCáOS 2 fonts!
6.9
Iconsprites <Obey$Dir>.!Sprites
6.9
RMEnsure FontManager 2.89 Obey
6.9
| The next command boots up your
6.9
| normal font directory
6.9
FontInstall <Obey$Dir>.
6.9
SetEval MoreFonts 0
6.9
The !Fonts.!Run file should be amended as follows:
6.9
| !Run file for !Fonts,
6.9
| version 1.05 (20-Aug-91)
6.9
RMEnsure FontManager 2.89 Error These fonts are for use with
6.9
RISCáOS 2.12 or later
6.9
|Obey <Obey$Dir>.!Boot
6.9
| The extra *FontInstall means that,
6.9
| if you double-click the Fonts
6.9
| directory, all such directories
6.9
| will be re-searched for any new
6.9
| fonts.
6.9
SetEval MoreFonts MoreFonts EOR 1
6.9
IF MoreFonts=1 THEN FontInstall <Obey$Dir>. ELSE FontRemove <Obey$Dir>.
6.9
|FontInstall
6.9
The !Boot file is only run on the first time that the filer sees the
!Fonts application, hence the initialisation of MoreFonts to zero. Every
time the directory is double-clicked, the !Run file will toggle the
value of this variable between 0 and 1. This is done in the line:
6.9
SetEval MoreFonts MoreFonts EOR 1
6.9
If the value is 1, the FontInstall command adds the new directory to the
font path and initialises all of the fonts. If the value is 0, the
FontRemove command removes the temporary font directory from the font
path.
6.9
If you require more than one spare directory, you will need to use
different variables, such as MoreFonts2 or FancyFonts.
6.9
In practice, it is better to have a directory called fonts inside your
!Fonts application which contains your preferred normal fonts at switch-
on, and a directory called spare (for example) which contains the set of
fonts that you wish to switch in and out. Then in your !Boot file, the
FontInstall command becomes
6.9
FontInstall <Obey$Dir>.fonts.
6.9
and the IF...THEN statement in the !Run file has <Obey$Dir>.spare. in
place of the two occurrences of <Obey$Dir>. .
6.9
Miscellaneous Comments
6.9
FontEd
6.9
When Tom Rank spoke to an Acorn representative he was told that they
were not contemplating the release of FontEd and that it is NOT public
domain. An early version got there by mistake via SID and it really
ought not, he said, to be distributed by PD libraries.
6.9
The Acorn representative acknowledged that Acorn were unlikely to do
anything about this situation, as it was öonly an early version and
wonæt work on RISCáOS 3ò. According to Tom this is partially true; a
friend of his has had problems with version 0.28, but at least some of
the time it runs on an A5000. However, he observes that it is a pity
that Acorn actually seem to be discouraging anyone else from working on
fonts. (There is a version of FontEd on Careware 7, released in May
1990, but it has been updated since then. Acorn havenæt objected,
presumably because of the charitable aspect. Ed.)
6.9
Screen blanking during floppy disc access
6.9
Stuart Bell writes: öMany things are very much more elegant in RISCáOS 3
but with one notable exception. I couldnæt believe my eyes when I saw
what happens to the screen during floppy disc accessing. Whether my ARM3
cache is dis- or enabled, in mode 24 the screen either flashes or is
blanked when loading, saving or formatting. If RISCáOS 2 didnæt need to
do this, why is it allowed now? (It did, in fact, happen in mode 21
under RISCáOSá2 ù HE.) Some re-coding is needed, even if the resultant
code is a little slower. More significantly, who was the Acorn person in
authority who thought that it is acceptable in 1993 to produce a
computer which professes to be a Éworld-beateræ which, when using floppy
discs, looks like the Sinclair Spectrum (or was it the ZX81?) reading
from cassette tape circa 1983? This just looks so tacky. I know that
serious use demands a hard disc, but first impressions count for a lot,
and a lot of demonstrations will be on floppy-based A3010æs. This is not
good enough, Acorn.ò
6.9
Basic program compaction errors
6.9
R. W. Darlington writes: öI am writing on a matter of utmost urgency.
6.9
öThere seems to be a growing urge amongst programmers to use a program
compactor on Basic RunImage files. Unfortunately, I know of no Basic
compactor program that does not exhibit a certain disastrous fault.
6.9
öIf, for instance, a Basic program contains any floating point or pseudo
variables followed by a Basic keyword or statement, then when it is
compacted and the space between the variable and the keyword is removed,
that line will generate an error.
6.9
öFor instance, if the program contained the line
6.9
10 IF TRUE THEN VDU7
6.9
which when compacted becomes
6.9
10IFTRUETHENVDU7
6.9
the lack of space between TRUE and THEN will generate an error.
6.9
öThe same is true for any line containing a floating point variable:
6.9
10 ha=TRUE
6.9
20 IF ha THEN VDU7
6.9
will crash when compacted because of a lack of space between ha and
THEN.
6.9
öI have come across numerous examples of RunImage files which have these
errors in them. One of them is Acornæs new RISCáOS 3 !SciCalc. Others
are new versions of Translator and Creator.
6.9
öPlease can you draw Archive readersæ attention to these problems as
soon as possible? Or, if any reader knows of a Basic compactor that does
not produce these errors, please can they let us know where it may be
obtained?ò
6.9
Matters Arising
6.9
Setting Caps Lock (*FX 202)
6.9
Unfortunately, the advice I gave last month about using *FX 202 may have
been a bit confusing because, after explaining that Peter Young wanted a
command to turn Caps Lock on, I then described how to turn it off! The
command to turn Caps Lock on, leaving the other settings unchanged, is
*FXá202,0,239.
6.9
The shortened form *FXá202,32 should turn Caps Lock on, but may have
side effects.
6.9
Note also that, to make sure that the LED indicators are correctly
updated after a *FXá202 command, it is a good idea to follow it with the
simple command *FXá118.
6.9
The bits of the keyboard status byte have the following meanings:
6.9
bit value meaning
6.9
0 (1) 0 Alt off
6.9
1 Alt on
6.9
1 (2) 0 Scroll Lock off
6.9
1 Scroll Lock on
6.9
2 (4) 0 Num Lock on
6.9
1 Num Lock off
6.9
3 (8) 0 Shift off
6.9
1 Shift on
6.9
4 (16) 0 Caps Lock on
6.9
1 Caps Lock off
6.9
5 (32) (öNormally setò)
6.9
6 (64) 0 Ctrl off
6.9
1 Ctrl on
6.9
7 (128) 0 Shift enable off
6.9
1 Shift enable on
6.9
(This information is taken from the RISC¡OSá2 Programmersæ Reference
Manual.)
6.9
Boot-up options (Archive 6.8 p26)
6.9
Peter Young has suggested that it might be better to use INKEY$ with a
delay of, say, 5 to 10 seconds instead of GET$ so that, if nobody
presses a key, the computer will continue with the boot file after a
short delay.
6.9
Unexpected messages from IDE filing system (Archive 6.8 p24)
6.9
Peter Young (who has a Beebug IDE interface) has confirmed that, even
after creating space for a directory cache, he still gets unexpected
messages unless he uses full pathnames. For instance, if he starts with
the commands
6.9
*DIR IDEFS::IDEDisc4.$.BASIC.B
6.9
*CAT
6.9
he gets a catalogue of the subdirectory ÉBæ as expected. However, if he
uses the abbreviated pathname instead:
6.9
*DIR $.BASIC.B
6.9
*CAT
6.9
he gets a catalogue of the directory ÉBASICæ even though the heading of
the directory listing correctly says that the CSD is IDEFS::IDEDisc4.$.
BASIC.B. The computeræs confusion is further demonstrated by the fact
that if he then enters:
6.9
*DIR ^
6.9
*CAT
6.9
it replies Directory É@æ not found.
6.9
ÉInterruptæ module (Archive 6.6 p55)
6.9
In March, we reported M. P. Sawleæs request for help in getting the
Interrupt module (Archive Shareware disc 6) working under RISCáOSá3, but
since then neither he nor we have heard anything. Does no-one else use
this module? Can its author help?
6.9
Using RISCáOSá3.10
6.8
Hugh Eagle
6.8
Encouragement for the faint-hearted!
6.8
Peter Young says that he is not only electronically illiterate but also
notoriously clumsy and so he left fitting the upgrade for over three
months. When he finally did dare to do it, he found it a lot easier than
he had dreaded Ö he broke nothing and it all worked first time. The only
worrying bit was changing the links, because the diagram was not clear
but a quick call to the suppliers (Beebug) soon put him right. So his
advice for others is: donæt be scared!
6.8
RISC OSá3.11
6.8
Newer machines and upgrades are now being shipped with version 3.11 of
the operating system. Barry Thompson says that this clears up a NetUtils
problem which occurs when computers equipped with RISCáOSá3.10 are used
on a network. The problem was that if you opened or extended a file
longer than 64Kb, anything beyond 64Kb was filled with null characters.
Existing network users can obtain a module (NetUtils) from their local
Acorn dealer.
6.8
Comparing the 3.11 and 3.10 Release Notes, the only other differences I
could find were the following brief notes:
6.8
Å DOS partition size ù DOSFS cannot be used with DOS partitions of 32Mb
or larger.
6.8
Å Acorn Lisp ù you must issue a *FXá13,11 command before loading Acorn
LISP.
6.8
Å Inspiration ù versions of Inspiration up to 1.03 will not run with
RISCáOSá3.10 or later.
6.8
(I gather that 3.11 also improves the situation with the ADFSbuffers
problem. On a 1Mb machine with 3.10, you are still advised to set
ADFSbuffers to 0. This is no longer necessary with 3.11. Ed.)
6.8
Programs that work
6.8
Mah Jong, The Game: Peter Young reports that the problem he mentioned
last month was caused by an inappropriate *FX call which he had inserted
in the !Run file ù so there seems to be nothing wrong with Mah Jong
itself.
6.8
Hard Disc Companion: Peter Young says that setting an IDEFS cache (see
below) has speeded this up considerably.
6.8
Ovation: In Archive 6.3 p14, we reported a problem with the printing of
indented lines from Ovation. Ken Cowap has now received a new version of
the program (version 1.37S) which, amongst other things, specifically
cures this particular problem. He says the problem turned out to be due
not to Ovation but to the RISCáOSá3 printer drivers. Although Acorn
acknowledged this, they saw no immediate possibility of a fix, and Risc
Developments very commendably undertook to rewrite part of Ovation.
6.8
Program problems
6.8
Aldebaran: Atle Mjelde Bσrdholt says that you must quit all application
tasks first (enter the Task display and Quit all tasks including
Pinboard), then run the game and it will install itself on the iconbar.
He says another solution is to exit the desktop and then type
*!Aldebaran at the command prompt. (Surely, Aldebaran is one of the most
recent games. Can it be true that it doesnæt run from the desktop?)
6.8
Atle Mjelde Bσrdholt also says that old Minerva games generally
reconfigure your machine. To avoid this, skip the !Boot file and click
on the next file in the sequence (e.g. in Hoverbod BS1 and in Missile
Control Title).
6.8
Ralph Gibbons finds that Break147 doesnæt work and would like to know if
there is a fix.
6.8
Mike Cook of Musbury Consultants repeats the point made by David Holden
last month: one reason older programs may not work is that the
abbreviations for some * commands have changed (e.g. *CHA. used to be
*ChannelVoice but is now *ChangeDynamicArea). Running the Obey file
Commands on the RISCáOSá3 support disc is supposed to help but if that
doesnæt, he suggests editing the program. This affects the PD sample
player !DSedit and his own !Sample.
6.8
Mike Cook also says that another change in the new operating system
resulted in the sampling part of his software not working. This also
affected the real time sound manipulator !Echo and the !Scope storage
oscilloscope. Any readers with the original programs can send him a disc
for a free upgrade. His address is 5áHelmshoreáRoad, Haslingden,
Rossendale, Lancashire, BB4á4BG.
6.8
Last month, we listed Powerband and Saloon Cars as not working. Rob
Brown says that Fourth Dimension can supply RISCáOSá3 compatible
replacements (for ú2 per program, he thinks).
6.8
Donæt forget that Rob Brown would welcome any contributions to his
compatibility chart, periodic updates of which will be included on the
Archive monthly program disc. His address is öValtaneeò, BrightonáRoad,
LoweráKingswood, Tadworth, Surrey, KT20á6UP.
6.8
Printing
6.8
öVDU 2ò printing
6.8
David Holden says that, although VDU 2 printing still works with
RISCáOSá3, it appears to monopolise the computer until printing is
finished. He has noticed that a couple of his programs that used to
multi-task, no longer do so and BBC emulator programs take a lot longer
to return control to the operator, even if using a printer buffer.
6.8
Printer buffering
6.8
Are the points in the previous paragraph perhaps connected to the non-
functioning of the RISCáOSá3 printer buffer mentioned last month? Or are
there further problems? Has anyone got a fix for it or them? The
questions seem endless!
6.8
Ralph Gibbons used to find a printer buffer routine (written by Anton
Calver and on Careware 5 & 7) very useful under RISCáOSá2. Does Anton
read Archive? Has he updated his routine to run under RISCáOSá3?
6.8
BJ330
6.8
In answer to Tord Erikssonæs plea (Archive 6.6 p56), Brian Cocksedge
recommends the Computer Concepts Turbo Driver (ú53 through Archive).
This produces much quicker printouts. The printer should be set to IBM
emulation (he understands this gives a better quality printout). The
stripes in graphics areas are a feature of the BJ printers due to minor
mechanical errors in the feed mechanism and the momentum of the print
head.
6.8
To print A3, as well as setting the paper size in the printer
preferences window, you have to tell the printer the size. Brian does
this by sending öEsc,C,99ò to the printer via a small Obey file
containing the commands:
6.8
*FX 138,27
6.8
*FX 138,67
6.8
*FX 138,99
6.8
LaserDirect
6.8
Steve Hutchinson experienced the same spurious lines that Roger King
reported in Archive 6.5 p31. He was told by CC that this was due to the
LaserDirect card not being happy in its podule slot. Rearranging podules
solved it for him.
6.8
Photocopying via LaserDirect
6.8
Tim Powys-Lybbe received version 2.09b of the LaserDirect software,
which cleared up the problems with print quality. However, his scanner
will now not do a photocopy at all, reporting (even with 3 Mb free) that
there is not enough free memory to do a photocopy. If he reverts to
version 2.05a, photocopying continues to work fine.
6.8
Disc drives
6.8
Speeding up an IDE drive
6.8
After he upgraded to RISCáOSá3, Peter Young initially found his IDE
drive (with a Beebug interface) to be very slow but eventually
discovered that setting an IDE cache speeded things up considerably. The
command to set it is:
6.8
*Configure IDEFSDirCache 16k
6.8
Before he created the cache, Peter was getting strange messages like dir
ö@ò not found when he specified directories at the command line with
references like $.xxx... rather than IDEFS:: IDEDisc4.$.xxx.... Since he
is now religiously using full pathnames, he doesnæt know whether
creating the cache might have solved the problem.
6.8
Peter says he still doesnæt really know whether what he wants is a cache
or a buffer. Can anyone throw any light on exactly what they do and what
the optimum sizes are for each in different circumstances? (Since
configuration options are provided, presumably one is expected to use
them, but when and how?)
6.8
SCSI discs
6.8
Andrew Shell encountered the same problems as Philip Lardner (Archive
6.6 p56) with copying, moving or deleting files to, from, or in the root
directory. By resetting his configuration so that *Status reads:
6.8
SCSIFSDirCache 16k
6.8
SCSFSdisks 1
6.8
SCSIFSDrive 4
6.8
File System SCSI
6.8
(and using !Configure to set öSCSI hard discsò toá1) he thought he had
solved the problem. All now seems to work correctly so far as saving,
copying and deleting are concerned. However, when he tried the Find
command on the !FONTS directory he got a sequence of finds reading
