Hi Knut

It's sure been a while... I do read all the posts that go through here but have
had very little time to play with the TRS-80's in the last year or so.

I do remember the discussions we had re a 16 bit interface.

I can't think why a 16 bit interface would make any difference to the drives
that are supported.  My original interface uses standard 16 bit IDE mode... it
doesn't run in the poorly supported IDE 8-bit mode.

Sectors on an IDE drive are 512 bytes (256 16 bit words).  The interface only
uses the bottom 8 bits of each word (the top 8 bits of each word are hard wired
to be 0) so to the TRS-80 sectors contain 256 bytes and the standard drivers are
happy.

This should work as long as there are no interface timing issues (it has always
been my suspicion that this was the issue with drives that did not work although
Mike's comment that the 74LS04 wasn't able to control EXTSEL is interesting).

I think it should also work with a Compact Flash card running in standard IDE
mode using one of the little compact flash adaptor boards.  I have one here so
will try and make some time to check it over over the Xmas break.

Apart from any timing design flaws in my interface, on reflection I think I
should have used the standard Hard Drive port numbers.  This would have allowed
the standard drivers to be used with minimum change.  I might investigate that
also.

Regards

Andrew


--- In TRS-80@yahoogroups.com, Knut <kr-lund@...> wrote:
>
> Hello Mike and Andrew
>
> Well, I have been thinking about this for a long time too. Discussing it in
the
> forums and with you Andrew before. I don't need an IDE drive or HD in general
as
> I'm fully satisfied with using 720K floppies... so it would be just for the
fun
> of doing it.
>
> I was thinking that I should rather make a 16bit interface for IDE so that we
> could use any IDE drive. I have some small flashdrives and a drive containing
a
> small CF card now so "the project" has got a bit more into focus.
>
> There is a GIDE project by Tilmann Reh general that does 16bit for Z80 and I
> thought I would do something similar and adapt your drivers Andrew.
>
> I think I have all the drivers from you Andrew also for LDOS 5.* I didn't
> remember but when I looked now it seemed to be so. I think I got those from
your
> site so they are probably still there? So I have never tried any of them.
>
> Knut
>
> jaquinn2001 wrote:
> > Hi Mike,
> >
> > Are you building my original design from
> > http://www.qsl.net/zl1wjq/trside1.htm
> > <http://www.qsl.net/zl1wjq/trside1.htm> ?
> >
> > I did start on LDOS 5.xx drivers but it was a long time ago. Give me a
> > few days to dig though the archives and see what I can find.
> >
> >  >From memory the main changes for LDOS 6 involved changing the port
> > addresses and adjustments to the "write product" check and some of the
> > error handling. I would assume that LDOS 5 would be similar.
> >
> > LDOS 6 was easier because the source was available from Frank Durda IV
> > at
> >
http://nemesis.lonestar.org/computers/tandy/software/os/logical_systems/lsdos6/s\
rc631/.
> >
<http://nemesis.lonestar.org/computers/tandy/software/os/logical_systems/lsdos6/\
src631/.>
> >
> > With 5 I think it is a patch the binary job.
> >
> > Regards
> >
> > Andrew
> >
> > --- In TRS-80@yahoogroups.com <mailto:TRS-80%40yahoogroups.com>, "mike"
> > <mikestv4@> wrote:
> >  >
> >  > has any one made ldos 5.xx drvers for this? got my working by ch out
> > u3 to a 7404(the 74ls04 do not have the balls to pull ext sel low)have
> > try to mod the hddvr6/dct and hdform6/cmd but do not have the
> > version/asm fill that makes it work...
> >  >
> >
> >
>

Hello,

This email message is a notification to let you know that
a file has been uploaded to the Files area of the TRS-80
group.

   File        : /CBtrs80.zip
   Uploaded by : redskulldc <layling@...>
   Description : CarteBlanche FPGA interface to Model-1

You can access this file at the URL:
http://groups.yahoo.com/group/TRS-80/files/CBtrs80.zip

To learn more about file sharing for your group, please visit:
http://help.yahoo.com/l/us/yahoo/groups/original/general.htmlfiles

Regards,

redskulldc <layling@...>

Hi All,

CBtrs80 interface for the Model-1 now supports Larry Campanell's Model-1 version
of Andrew Quinn's IDE interface.

CBtrs80.zip in the files section for details.

Has been tested with a few Seagate IDE drives, responds as expected to Status,
read/write requests.
Not tested with CF cards as I don't have any here at present :0)

Next on the list will be VGA display and MicroLABS (model3) Hires functionality.

Cheers,
Red

Hi Andrew, et al.

Don't suppose that one of you guys with a working HD using this IDE interface
could possibly make a HD disk image and upload it?

Easy under Linux, or Ciderpress for the Apple2 can make images of Foreign HD's
under Windoze.

Cheers,
Leslie


--- In TRS-80@yahoogroups.com, "jaquinn2001" <jaquinn@...> wrote:
>
> Hi Mike,
>
> Are you building my original design from http://www.qsl.net/zl1wjq/trside1.htm
?
>
> I did start on LDOS 5.xx drivers but it was a long time ago.  Give me a few
days to dig though the archives and see what I can find.
>
> From memory the main changes for LDOS 6 involved changing the port addresses
and adjustments to the "write product" check and some of the error handling.  I
would assume that LDOS 5 would be similar.
..
> Regards
>
> Andrew

jaquinn2001 wrote:
> Hi Knut
>
> It's sure been a while... I do read all the posts that go through here but
> have had very little time to play with the TRS-80's in the last year or so.
>
> I do remember the discussions we had re a 16 bit interface.
>
> I can't think why a 16 bit interface would make any difference to the drives
> that are supported.  My original interface uses standard 16 bit IDE mode...
> it doesn't run in the poorly supported IDE 8-bit mode.

Ok, I remembered wrong or made bad assumptions, sorry. So this is not why the
interface is picky on drives.

> Sectors on an IDE drive are 512 bytes (256 16 bit words).  The interface only
> uses the bottom 8 bits of each word (the top 8 bits of each word are hard
> wired to be 0) so to the TRS-80 sectors contain 256 bytes and the standard
> drivers are happy.

Yes, we need 256 byte sectors for the DOS so this is a good way to do it, I
agree. I guess you would need to buffer in the driver if the high byte part of
the sector should be used (256+ bytes to the driver).

> This should work as long as there are no interface timing issues (it has
> always been my suspicion that this was the issue with drives that did not
> work although Mike's comment that the 74LS04 wasn't able to control EXTSEL is
> interesting).

Yes, I can see it might be a problem. It is used both for EXTIOSEL and /IOR.
Model III/4 does pullup this pin with 150 ohm. IDE's /IOR seems not to be
terminated though. Maybe the pullup lifts the signal too much for the IDE /IOR
to recognize it as low. I think maybe it would be better to split those and
drive EXTIOSEL with an open collector driver. This is simpler for model-I and
the EACA clones as there is no EXTIOSEL signal.

> I think it should also work with a Compact Flash card running in standard IDE
> mode using one of the little compact flash adaptor boards.  I have one here
> so will try and make some time to check it over over the Xmas break.

The CF thingy I have is a very slim drive with a CF card hidden inside. But I
also have several IDE flash drives now that are not based on CF. For example
32MB 2.5".

> Apart from any timing design flaws in my interface, on reflection I think I
> should have used the standard Hard Drive port numbers.  This would have
> allowed the standard drivers to be used with minimum change.  I might
> investigate that also.

Yes, good that more people are interested.

Just now I'm more focused on Silicon Graphics stuff. (This is hobby and TRS-80
and compatibles aren't the only computers here, though they are my favorites.) I
did manage to interface some asus stereo glasses to the Silicon Graphics
machines (Indigo2 and Indy) so I do spectacular 3D stuff on these supercomputers
of the 90s. :-)

Maybe IDE on the Video Genie will get priority soon.

Knut

Hi Leslie,

Can you explain why you are interested in an image?

Unfortunately the TRS-80 doesn't boot directly from the hard drive so you still
need a boot floppy that loads the hard disk drivers.  An image of a drive
wouldn't be a lot of use unless you have the drivers loaded.

... but perhaps I missed the reason for your questions.

Regards

Andrew

--- In TRS-80@yahoogroups.com, "redskulldc" <layling@...> wrote:
>
> Hi Andrew, et al.
>
> Don't suppose that one of you guys with a working HD using this IDE interface
could possibly make a HD disk image and upload it?
>
> Easy under Linux, or Ciderpress for the Apple2 can make images of Foreign HD's
under Windoze.
>
> Cheers,
> Leslie
>
>
> --- In TRS-80@yahoogroups.com, "jaquinn2001" <jaquinn@> wrote:
> >
> > Hi Mike,
> >
> > Are you building my original design from
http://www.qsl.net/zl1wjq/trside1.htm ?
> >
> > I did start on LDOS 5.xx drivers but it was a long time ago.  Give me a few
days to dig though the archives and see what I can find.
> >
> > From memory the main changes for LDOS 6 involved changing the port addresses
and adjustments to the "write product" check and some of the error handling.  I
would assume that LDOS 5 would be similar.
> ..
> > Regards
> >
> > Andrew
>

Had lunch with Don French today. He is doing well. We had a great discussion of
the old days when I hired him as a young man in High School to work in the La
Mesa Radio Shack. Glad to find a group like this that still has an interest in
the beginnings of the personal computer business.

Hi Andrew, et al.

> This should work as long as there are no interface timing issues
>....
> I think it should also work with a Compact Flash card running in
> standard IDE mode using one of the little compact flash adaptor
> boards.  I have one here so will try and make some time to check it
> over over the Xmas break.

Design as it is doesn't meet the timing specs for CF cards.
Address lines need to be stable and Select asserted for a minimum of 60ns before
RD or WR are asserted.

"Real" HD's (that I have tried anyway) are much more lax in regards to this
requirement, and generally don't seem to mind if all signals are asserted at the
same time.
This might suggest that CF cards are clocked internally at a much slower rate
than the logic on HD's.

---

As far as the HD image is concerned, I've been playing around with the
CarteBlanche FPGA card (as you may have seen on this forum), and recently added
the functionality for your IDE interface.
Didn't even cross my mind that the Level2 ROMS don't have this functionality!
hehe
Using the CB card it would be easy to overlay the internal ROMS with a modified
Level2 which had HD boot support.

Cheers,
Leslie

Hi Leslie,

Thanks for the comment on the timing.  Do you have any recommendations on an
efficient way to add the delay on the RD and WR lines.

The ideal would be to keep the design using 74xx type parts which would be
available to anyone rather than requiring an FPGA or CPLD.

Regards

Andrew

--- In TRS-80@yahoogroups.com, "redskulldc" <layling@...> wrote:
>
> Hi Andrew, et al.
>
> > This should work as long as there are no interface timing issues
> >....
> > I think it should also work with a Compact Flash card running in
> > standard IDE mode using one of the little compact flash adaptor
> > boards.  I have one here so will try and make some time to check it
> > over over the Xmas break.
>
> Design as it is doesn't meet the timing specs for CF cards.
> Address lines need to be stable and Select asserted for a minimum of 60ns
before RD or WR are asserted.
>
> "Real" HD's (that I have tried anyway) are much more lax in regards to this
requirement, and generally don't seem to mind if all signals are asserted at the
same time.
> This might suggest that CF cards are clocked internally at a much slower rate
than the logic on HD's.
>
> ---
>
> As far as the HD image is concerned, I've been playing around with the
CarteBlanche FPGA card (as you may have seen on this forum), and recently added
the functionality for your IDE interface.
> Didn't even cross my mind that the Level2 ROMS don't have this functionality!
hehe
> Using the CB card it would be easy to overlay the internal ROMS with a
modified Level2 which had HD boot support.
>
> Cheers,
> Leslie
>

--- In TRS-80@yahoogroups.com, "Neil Morrison" <neilsmorr@...> wrote:
>
> Here are the first 10 of ThumpNugget's scans of Softside  magazine.
>
> If you don't understand Rapidshare, always select the 'Free' options.
>
> I'll post the rest as time permits.

Neil,

I tried, and selected the "Free" option, and was informed the file could only be
downloaded if you have a premium account.

Howard

On 21 Nov 2009 at 13:43, howipepper wrote:

> --- In TRS-80@yahoogroups.com, "Neil Morrison" <neilsmorr@...> wrote:
> >
> > Here are the first 10 of ThumpNugget's scans of Softside  magazine.
> >
> > If you don't understand Rapidshare, always select the 'Free' options.
> >
> > I'll post the rest as time permits.
>
> Neil,
>
> I tried, and selected the "Free" option, and was informed the file could only
be downloaded if you have a premium account.

If you read carefully, you'll find that you were told that there
weren't any free slots available *at that time*.

If you try again in an hour or so, you'll usually get a shot at it.

--
Leonard Erickson (aka shadow)
shadow at shadowgard dot com

is thare a book or a list of the jumps for this dos need the mod 3's...

http://rapidshare.com/files/309373880/SoftSide_11_Vol_1-11_1979-08_Meltdown.pdf
Size: 8746

http://rapidshare.com/files/309373882/SoftSide_12_Vol_1-12_1979-09_Engineer.pdf
Size: 5957 KB

http://rapidshare.com/files/309373883/SoftSide_13_Vol_2-01_1979-12_Westward.pdf
Size: 19793 KB

http://rapidshare.com/files/309373884/SoftSide_14_Vol_2-02_1979-11_Mount_Everest\
.pdf
Size: 7691 KB

http://rapidshare.com/files/309373888/SoftSide_15_Vol_2-03_1979-12_Oil_Baron.pdf
Size: 6915 KB

http://rapidshare.com/files/309373889/SoftSide_16_Vol_2-04_1980-01_Moving_Maze.p\
df
Size: 5113 KB

http://rapidshare.com/files/309373890/SoftSide_17_Vol_2-05_1980-02_Land_the_Shut\
tle.pdf
Size: 7664 KB

http://rapidshare.com/files/309373892/SoftSide_18_Vol_2-06_1980-03_Produce_a_Bro\
adway_Play.pdf
Size: 7417 KB

http://rapidshare.com/files/309373894/SoftSide_19_Vol_2-07_1980-04_Casino_Games.\
pdf
Size: 7031 KB

http://rapidshare.com/files/309373895/SoftSide_20_Vol_2-08_1980-05_Star_Trek.pdf
Size: 13768 KB

mike wrote:
> is thare a book or a list of the jumps for this dos need the mod 3's...
>
The model I/III DOSPLUS manual is very complete I don't know if it is available
for download anywhere but... I seem to remember someone had ocr'ed it but I
couldn't find that I have it.

note that some functions does not have the same functionality or parameters as
LDOS, one indicator for this is when the address differs.

I just ocr'ed it so I hope all is correct, did skim through the M3 addresses...

Knut

DOSPLUS User's Manual vA.3 Page - T/23

Routine name    Address         Page
Name    Model I Model III       Number
ABORT@  4030H   4030H   T/30
BKSP@   4445H   4445H   T/28
CAT@    445AH   4419H   T/33
CKEOF@  4451H   4457H   T/29
CLOSE@  4428H   4428H   T/25
CMD@    4400H   4400H   T/32
CMNDI@  4405H   4405H   T/32
DEBUG@  440DH   440DH   T/32
DISKIO% 4485H   4488H   T/44
DIVD@   444EH   4451H   T/36
DSP@    0033H   0033H   T/49
DSPLY@  4467H   4467H   T/38
ERROR@  4409H   4409H   T/32
EVAL@   4479H   4479H   T/39
EXIT@   402DH   402DH   T/30
FEXT@   4473H   444BH   T/29
FILPTR@ 4454H   428DH   T/30
FSPEC@  441CH   441CH   T/24
GET@    0013H   0013H   T/48
GTDATE@ 4470H   4470H   T/39
GTTIME@ 446DH   446DH   T/38
INIT@   4420H   4420H   T/24
KBD@    002BH   002BH   T/48
KEY@    0049H   0049H   T/49
KEYIN@  0040H   0040H   T/49
KILL@   442CH   442CH   T/25
LOAD@   4430H   4430H   T/26
LOCDCB% 4488H   44A0H   T/46
LOCDCT% 448BH   44A3H   T/47
LOCDEV@ 447FH   447FH   T/42
MULT@   444BH   444EH   T/36
OPEN@   4424H   4424H   T/25
PARAM@  4476H   4454H   T/37
PEOF@   4448H   4448H   T/28
POSN@   4442H   4442H   T/28
PRINT@  446AH   446AH   T/38
PRT@    003BH   003BH   T/49
PUT@    001BH   001BH   T/48
RAMDIR@ 4457H   4290H   T/31
READ@   4436H   4436H   T/26
RES @   4413H   4416H   T/35
REW@    443FH   443FH   T/27
RUN@    4433H   4433H   T/26
SET@    4410H   4413H   T/34
SORT@   4482H   4482H   T/43
VER@    443CH   443CH   T/27
WILD@   447CH   447CH   T/42
WRITE@  4439H   4439H   T/27

http://rapidshare.com/files/309397926/SoftSide_21_Vol_2-09_1980-06_Micro_Million\
aire.pdf
Size:9435 KB

http://rapidshare.com/files/309398326/SoftSide_22_Vol_2-10_1980-07_Adventureland\
.pdf
Size:6617 KB

http://rapidshare.com/files/309401800/SoftSide_23_Vol_2-11_1980-08_Fly_a_Sailpla\
ne.pdf
Size:57793 KB

--- In TRS-80@yahoogroups.com, "bartlett.peterj" <bartlett.peterj@...> wrote:
>
> Hello,
>
> I'm thinking about recreating the "VOX Box" peripheral for the Model I that
Radio Shack sold way-back-when.  I have the service manual with the schematic,
and it's pretty simple: just a bunch of amplifiers on the microphone signal and
a jury-rigged ADC.  The software does most of the processing.
>
> But I have a couple questions:
> 1. Does it work?  Can anyone who had or has one comment on how well it
recognizes speech?
>
> 2. Does anyone have the original software?  I have the files from Ira's old
web site called "VOX Box Program Source Code (19xx)(-)[SRC]", but it's
decompiled assembly and doesn't seem to go with the diagnostic software in the
service manual, so I think it's the wrong software.
>
> Any guidance and/or help would be appreciated.
>
> Pete.
>

An update, since the group has been pretty quiet...

I've put together a prototype VoxBox, and it works pretty well.  It recognizes
words with about 90% accuracy.  It has the most trouble with words that start
with high frequency sound like "six" or "seven".  I think with a little more
tweaking of the filtering I can get it to work even better... as well as such a
simple implementation can work, anyway.

By using a PC-style electret (voltage biased) microphone instead of the original
CB-style "10-4, good buddy" passive mic, I was able to completely remove one of
the two input amplifier stages.

A word of warning for anyone building one of these... the service manual lists
R17 as 150 ohms, 1/4 watt.  With all voltages nominal its power consumption is
just over 1/4 watt, and with just a little power supply variance it could be
nearly 1/2 watt... use at least 1/2 watt resistor.  Also, the 1N752 zener diode
CR1 runs pretty hot, near its max specified current; I'm going to find one that
can handle more reverse current at about the same zener voltage before I build
my final device.

As far as software, I assembled the old version of the driver from Ira's web
site, and re-disassembled it using a much better tool.  It originally used the
4K chunk of memory from F000-FFFF, but it was easily relocated to E000-EFFF
since I use LDOS which puts the double-density and hard disk drivers in high
memory.  Also, for some reason the original VoxBox driver didn't disable
interrupts while training or recognizing words, which seem like a fairly
time-sensistive algorithms.  I fixed that.

For application software, there are several BASIC diagnostic program listings in
the service manual, and several application program listings in the user manual
including an inventory app, a lunar lander game, and a voice-directed drawing
program.  Big fun!

If anyone else out there builds one of these devices, I'd be interested in
hearing how it goes.

Pete.

http://rapidshare.com/files/309404922/SoftSide_24_Vol_2-12_1980-09_Dr_Livingston\
.pdf
Size:51684 KB

http://rapidshare.com/files/309408469/SoftSide_25_Vol_3-01_1980-10_Earth_Port_II\
.pdf
Size:61457

http://rapidshare.com/files/309412908/SoftSide_26_Vol_3-02_1980-11_Kriegspiel.pd\
f
Size:68041 KB

Neil

I've been unable to get the Softsides from Rapidshare. The whole point of the
free side of Rapidshare is to frustrate you, so you'll pay for it.

Can you please make torrents of the files? Or, post them in the files section
here?

At 5:07 PM +0000 11/25/09, Mage wrote:
>I've been unable to get the Softsides from Rapidshare. The whole
>point of the free side of Rapidshare is to frustrate you, so you'll
>pay for it.
>
>Can you please make torrents of the files? Or, post them in the
>files section here?

There is already a torrent file for this.  It's 2.77GB and contains
most issues through early '84.  It's 2.77GB total.  Use any torrent
search engine like http://isohunt.com and search for Softside.  You
should find it.  Please make sure you share what you download.
Bittorrent works only because people continue to share the files.

No disrespect to Neil for posting them, not everyone wants to use
Bittorrent, so thanks Neil!  But if you haven't tried Bittorrent, now
is the time to try it.  Downloading all these files from Rapidshare
is very painful.  Get a copy of Vuze, or uTorrent.  It might be a
little confusing to set up due to firewall issues, but there should
be an FAQ, Wiki, or other "new user" set up page to help you through
it.

I've been serving several vintage computer torrents all year
including TRS-80 documents, manuals, etc. (a 17GB torrent download),
and some software for the color computer.  I have an archive of
TRS-80 M1, III, 4/4P Software that I was going to post eventually,
but I need to sort through it first.  Again, a large archive that I
can only serve via Bittorrent.  Serving this on a web site is just
out of the question due to shear quantity of material.


Regards,
Mike
--
----------------
Michael Lohmeyer
lohmeyer@...

Download a copy of USDownloader or (and better) JDownloader. Either will
automate the process for you - even while you sleep. I've downloaded 30 or 40
files that way without touching the keyboard.

I can't post them here - they're too big. I may post the listings if I can
convert them to text.

Neil

--- In TRS-80@yahoogroups.com, "Mage" <alhartman6@...> wrote:
>
> I've been unable to get the Softsides from Rapidshare. The whole point of the
free side of Rapidshare is to frustrate you, so you'll pay for it.
>
> Can you please make torrents of the files? Or, post them in the files section
here?
>

Watch out. There are two torrents. Only the smaller one works - the other was a
mistake but there's no way to kill it.

Neil

--- In TRS-80@yahoogroups.com, Michael Lohmeyer <lohmeyer@...> wrote

> There is already a torrent file for this.  It's 2.77GB and contains
> most issues through early '84.  It's 2.77GB total.  Use any torrent
> search engine like http://isohunt.com and search for Softside.

http://rapidshare.com/files/309417212/SoftSide_27_Vol_3-03_1980-12_Kidnapped.pdf
Size:58683 KB

http://rapidshare.com/files/309421143/SoftSide_28_Vol_3-04_1981-01_Convoy.pdf
Size:61544 KB

http://rapidshare.com/files/309424921/SoftSide_29_Vol_3-05_1981-02_Miner.pdf
Size:62139 KB

Neil

--- In TRS-80@yahoogroups.com, Mark McDougall <msmcdoug@...> wrote:
>
> skipp025 wrote:
>
> > "The author of NEWDOS/80 was always very insistent on
> > maintaining his privacy, and didn't even want his name
> > to be made public."
>
> Well that's perfectly understandable, given the massive public controversy
> over NEWDOS/80 at the time; the childhood safety concerns, the impact on the
> environment, the number of people that would be losing their jobs because of
> it, how much cheaper it would be to make and sell illicit drugs because of
> it... ????
>
> Regards,
>
> --
> |              Mark McDougall                | "Electrical Engineers do it
> |  <http://members.iinet.net.au/~msmcdoug>   |   with less resistance!"
>


Sarcasm aside (although I did chuckle), I was trying to figure out why the
author would want to remain anonymous.

I wonder if at the time there where any legal threat from Tandy or if any of the
code in NEWDOS/80 is re-used from NEWDOS or other places.  I can't think of why
else -- i.e. *some legal issues* he'd want to keep under the radar.   Please
note that I am *not* concluding that the author did do anything that would be
civilly actionable, just that if the environment was litigious I could totally
understand why they would not want to be known.

Also, I wonder if it's possible the author had a conflict of interest by somehow
working for another company that was competing directly or indirectly.

There are many ways to speculate, and unfortunately, many ways to presume some
wrongness on the authors part which isn't really fair.

<sarcasm>
Until I know for sure I will assume Bill Gates wrote NEWDOS/80.  Clearly we can
see why he wouldn't want to be exposed.

What could be more humiliating than being associated with Windows Vista, the
operating system that makes 'Bob' look good?

Neil

--- In TRS-80@yahoogroups.com, "David S" <turnkit@...> wrote:

> Until I know for sure I will assume Bill Gates wrote NEWDOS/80.  Clearly we
can see why he wouldn't want to be exposed.