Testing the calculator on my eTrex Legend with firmware v2.40c

August 15, 2002

I recently acquired an eTrex Legend, and shortly thereafter upgraded the firmware from the factory shipped version (2.38) to the latest beta version (2.40c). One of the claimed improvements of the new version was improved calculator function. I didn't play with the calculator much BEFORE upgrading, so I can't compare, but I did find some interesting (and a few just plain wrong) results:

-the calculator displays up to 6 places past the decimal, but the user can only enter numbers with up to 5 digits past the decimal

-EXCEPTION to above: if the number begins with all 0's numbers as small as .0000000001 can be entered, but most calculations involving it will lose it (due to the precision of the displayed results) however you can get results from the 1/X and ln functions on these very small numbers.

-entry of numbers with multiple decimal points is allowed (up to 3 decimal points), however numbers after the 2nd decimal point seem to just be ignored--behaviour after entering/using these numbers can be odd until C or CE is pressed.

-calculator handles numbers as large as 9.899999 E+24 (yes, 9.899999, not 9.999999, anything over that gives "ERROR!") but you only directly enter a numbers up to 9999999999

-displays most(?) integers up to 99999999999999999 w/o scientific notation

-oddly, it handles -0 distinctly, but this does not seem to lead to any odd or incorrect results

-does not generate an error for divide by zero (gives result of 0 instead), or ln of zero and negative numbers (does nothing, displayed number does not change)

-results in scientific notation that should be E+10 display incorrectly as E+1, similarly results that should be E+20 displays as E+2 (proof: enter "4.3" and hit "X^2" repeatedly)

-algorithm for calculating tangent function a little off near and at the asymtote:

tan of 89.99999 gives result of "5729578.057851", should be 5729577.951308

tan of 89.999999 gives "57295790.748394", should be 57295779.5130823

tan of 90 degrees gives "-2.864114E+14" instead of an error

-repeated pressing of operation buttons which require 2 operands (+,-,*,/,X^Y) can lead some slightly unpredictable behavior

-would be nice if the selector wrapped around sides rather than going to the window-control button

-Sin, Cos and Tan of any number in scientific notation (as well as some very large numbers not in scientific notation) gives a result of "0" instead of the correct answer

-X^Y for some large-ish numbers gives result of "0" instead of "ERROR!" (e.g. 256^256)

df

[This message was edited by DarrenF on August 15, 2002 at 02:24 PM.]

August 15, 2002

Have you contacted Garmin yet?

August 15, 2002

Okay, I couldn't resist. I once wrote a floating point math library, and the package to present a decimal interface with the user. I can attest that this isn't the sort of thing one does lightly. It is really tricky to get it all just right. There are zillions of gotchas.

Guess some programmer at Garmin didn't realize the land mine he was entering. Probably said, "Oh sure, I can knock that off in a week. Let's put it in the next release..."

2.3k · August 15, 2002

I have indeed reported som of these errors, like E+1 instead of E+10, and some of the math errors, to Garmin. Now, that's for the Vista calculator, but they are certainly identical, so that doesn't matter.

Just WHY is there no "inv" key to the trig and log functions???

The findings by DarrenF above makes me wonder what they did fix in this update?

Anders

[This message was edited by Anders on August 16, 2002 at 02:40 AM.]

August 16, 2002

st_richardson: No, I haven't contacted Garmin. I do not know anyone at Garmin, and from what I've seen posted on this message board it seems there is very little likelihood that an email sent to them regarding this matter would be read by someone who matters. Please forgive my pessimism. But if someone knows a Garmin email address other than the 'front door', that isn't read only by imbeciles who assume they're answering the dumb questions of even-bigger-imbecile-users, I'll try that.

John, Mark & Kyle: Yes, the devil is in the details. Programming a calculator is is not nearly as simple as it sounds on the surface. I realize that this is just a "free accessory", not at all core to the function of the unit and I can understand if they don't really care about the minutia of its operation. But I'm of the opinion that if you're going to do something, you may as well do it right.

Anders: Yes, shortly after I posed my findings, I searched the arhives and ran across your prior post about the missing exponent digit and ln incorrect function. I need to make it a habit to do that sort of search BEFORE posting . I'm not sure what they fixed since, as I said, I neglected to give the calculator in 2.38 a good once-over before I upgraded. What does the Vista's calculator do for ln on a negative number? (give 0, error, leave the displayed number unchanged, or other). Seems like I read that the C and CE buttons weren't working right before; they seem to function normally from what I can tell.

I agree with you about the INV key. I would trade the "%" key on the scientific calc for it. There would be no need to squeeze more letters onto the keys, the displayed labels could simply change from Tan, Cos, Sin, ln to Atan, ACos, Asin, e^x as the INV key was toggled. If they really wanted to get fancy, X^Y, X^2, (root)X could become X(root)Y, LOG, 10^X...

I made another observation: in most in-fix calculators (although I'd love to see a RPN calc included in the future...) pressing two operation buttons in a row usually causes the first one to be ignored and the 2nd one to apply. For example, if you press "3 +", and then realize "Crap, I mean to multiply, not add" and press "* 3 =", you get 9 (3*3) on well-programmed calculators. The eTrex calculator doesn't function this way; pressing the second operation key causes the first operation to be executed with the displayed number as both operands (in the above example, the eTrex displays 6 when you hit * rather then ignore the + and await the 2nd operand for multiplication).

August 16, 2002

st_richardson: No, I haven't contacted Garmin. I do not know anyone at Garmin, and from what I've seen posted on this message board it seems there is very little likelihood that an email sent to them regarding this matter would be read by someone who matters. Please forgive my pessimism. But if someone knows a Garmin email address other than the 'front door', that isn't read only by imbeciles who assume they're answering the dumb questions of even-bigger-imbecile-users, I'll try that.

John, Mark & Kyle: Yes, the devil is in the details. Programming a calculator is is not nearly as simple as it sounds on the surface. I realize that this is just a "free accessory", not at all core to the function of the unit and I can understand if they don't really care about the minutia of its operation. But I'm of the opinion that if you're going to do something, you may as well do it right.

Anders: Yes, shortly after I posed my findings, I searched the arhives and ran across your prior post about the missing exponent digit and ln incorrect function. I need to make it a habit to do that sort of search BEFORE posting . I'm not sure what they fixed since, as I said, I neglected to give the calculator in 2.38 a good once-over before I upgraded. What does the Vista's calculator do for ln on a negative number? (give 0, error, leave the displayed number unchanged, or other). Seems like I read that the C and CE buttons weren't working right before; they seem to function normally from what I can tell.

I agree with you about the INV key. I would trade the "%" key on the scientific calc for it. There would be no need to squeeze more letters onto the keys, the displayed labels could simply change from Tan, Cos, Sin, ln to Atan, ACos, Asin, e^x as the INV key was toggled. If they really wanted to get fancy, X^Y, X^2, (root)X could become X(root)Y, LOG, 10^X...

I made another observation: in most in-fix calculators (although I'd love to see a RPN calc included in the future...) pressing two operation buttons in a row usually causes the first one to be ignored and the 2nd one to apply. For example, if you press "3 +", and then realize "Crap, I mean to multiply, not add" and press "* 3 =", you get 9 (3*3) on well-programmed calculators. The eTrex calculator doesn't function this way; pressing the second operation key causes the first operation to be executed with the displayed number as both operands (in the above example, the eTrex displays 6 when you hit * rather then ignore the + and await the 2nd operand for multiplication).

2.3k · August 19, 2002

I'm pretty sure the Vista and Legend calculators do work the same.

ln(-6) and ln(0) doesn't change the displayed number. sqrt(-6) gives the message "ERROR!", though, so it's not consequent.

ln(negative argument) gives a complex result, while ln(0) has an infinite result.

-3^4.3 also results in ERROR! (good, since the result is ~ 66+91i, and this calculator isn't supposed to do complex arithmetic), but -3^4 is correctly calculated as 81.

In the first release, using the "C" or "CE" keys, the number in the display didn't change until you started to key in the new number. No visual feedback of if you really cleared that erroneous entry before you started to key in the number again.

My old TI 59 (still works!) starts flashing the display (Error indication) if I press two operation keys without any number in between. My even older Commodore simply performs the last operation I press. Usually, only "you get if for free promoting calculators" behave as the one in the eTrex series.

Besides, I've told Garmin that I want the memory to retain its value, even when you leave the calculator. At least as long as the power is on. Cut and paste between the calculator and other numbers you can display would be nice, but I can't really see how they should implement that.

Anders

August 19, 2002

Good to see someone else with a high enough geek-quotient to be interested in the nuances of the calculator performance

I agree, consistancy should dictate that ln of zero or a negative operand as well as 1/X of zero should result in "ERROR!", just like square-root of a negative.

I'm glad to hear the C/CE issue, at least, has been rectified.

A Commodore calculator? I never knew they made calculators. Must have been a European thing--they were only known for computers over here (the venerable C-64)

I agree with you about that retaining the value in calculator memory (and perhaps even the value displayed) would be nice. But then I realized even the Windows calculator doesn't do that, then again, you can switch applications to do something else w/o closing it. Cut&Paste functionality does sound like a stretch, given the interface and limited inputs, however one simple approach would be to add a button to the on-screen keyboard which would paste the number from the calculator Memory into the current input field. Still no piece of cake to implement, though... I'd be happy with the hi-lighted button selection wrapping around the sides.

Oh, by the way, I discovered another calculation error: 0^0 gives a result of "ERROR!" instead of the correct result of "1".

I also played with the Calender a bit, but didn't find any problems. It correctly shows 2000 as a leap year. I did not wish to subject my clickstick to the ~1200 clicks required to check 1900 or 2100 (neither should be a leap year) much less 1600 or 2400 (both should be a leap year),

df

[This message was edited by DarrenF on August 19, 2002 at 02:45 PM.]

August 19, 2002

Good to see someone else with a high enough geek-quotient to be interested in the nuances of the calculator performance

I agree, consistancy should dictate that ln of zero or a negative operand as well as 1/X of zero should result in "ERROR!", just like square-root of a negative.

I'm glad to hear the C/CE issue, at least, has been rectified.

A Commodore calculator? I never knew they made calculators. Must have been a European thing--they were only known for computers over here (the venerable C-64)

I agree with you about that retaining the value in calculator memory (and perhaps even the value displayed) would be nice. But then I realized even the Windows calculator doesn't do that, then again, you can switch applications to do something else w/o closing it. Cut&Paste functionality does sound like a stretch, given the interface and limited inputs, however one simple approach would be to add a button to the on-screen keyboard which would paste the number from the calculator Memory into the current input field. Still no piece of cake to implement, though... I'd be happy with the hi-lighted button selection wrapping around the sides.

Oh, by the way, I discovered another calculation error: 0^0 gives a result of "ERROR!" instead of the correct result of "1".

I also played with the Calender a bit, but didn't find any problems. It correctly shows 2000 as a leap year. I did not wish to subject my clickstick to the ~1200 clicks required to check 1900 or 2100 (neither should be a leap year) much less 1600 or 2400 (both should be a leap year),

df

[This message was edited by DarrenF on August 19, 2002 at 02:45 PM.]

August 19, 2002

For DarrenF:

I think you should send your findings about the calculator to Garmin. You've done some good work. They may or may not be able to make the corrections -- the Legend, after all, was not designed to be a full-fledged scientific calculator, but it's worth a try.

I've sent a number of requests to their technical support and have always received a reasonably quick and helpful reply. The Etrex Legend Beta team's address is:

Yes, you've read some stinging criticisms of Garmin online but most, if not all of it, is just e-mail flaming.

August 19, 2002

For DarrenF:

I think you should send your findings about the calculator to Garmin. You've done some good work. They may or may not be able to make the corrections -- the Legend, after all, was not designed to be a full-fledged scientific calculator, but it's worth a try.

I've sent a number of requests to their technical support and have always received a reasonably quick and helpful reply. The Etrex Legend Beta team's address is:

Yes, you've read some stinging criticisms of Garmin online but most, if not all of it, is just e-mail flaming.

August 19, 2002

Thank you! I didn't realize there was an email address specifically for the purpose of reporting issues with the Legend's Beta firmware. I need to do more reading and less clicking of "Download Now"

I'll try to draft a well-written summary and email it to them.

The new scroll-by-page with the zoom-in/out buttons is very nice. However, it does not work on the following screens: "Find | Cities | By Name" and "Find | Points of Interest | By Name". Curiously, it works fine in "Find | Waypoints | By Name" (I didn't expect it to, thinking it was a problem with all the "By Name" dialog boxes.

Can anyone else with the new firmware confirm this? Can anyone find any other situations where scolling-by-page using the zoom buttons doesn't work?

The update notes indicate "Set zoom in and zoom out keys to page through SOME lists." [emphasis mine], but it appears to work on most lists (all I've encountered except the two listed above).

df

[This message was edited by DarrenF on August 19, 2002 at 02:43 PM.]

[This message was edited by DarrenF on August 20, 2002 at 07:30 AM.]

August 19, 2002

Thank you! I didn't realize there was an email address specifically for the purpose of reporting issues with the Legend's Beta firmware. I need to do more reading and less clicking of "Download Now"

I'll try to draft a well-written summary and email it to them.

The new scroll-by-page with the zoom-in/out buttons is very nice. However, it does not work on the following screens: "Find | Cities | By Name" and "Find | Points of Interest | By Name". Curiously, it works fine in "Find | Waypoints | By Name" (I didn't expect it to, thinking it was a problem with all the "By Name" dialog boxes.

Can anyone else with the new firmware confirm this? Can anyone find any other situations where scolling-by-page using the zoom buttons doesn't work?

The update notes indicate "Set zoom in and zoom out keys to page through SOME lists." [emphasis mine], but it appears to work on most lists (all I've encountered except the two listed above).

df

[This message was edited by DarrenF on August 19, 2002 at 02:43 PM.]

[This message was edited by DarrenF on August 20, 2002 at 07:30 AM.]

2.3k · August 19, 2002

There were several Commodore calculators for sale here, but that was before the Commodore 20 and 64 computers. We are talking 1975-1980 now. I still have a few of them in a box in my basement.

My Hewlett-Packard 48 SX says 0^0 is 1, but the HP 32 SII says that's an "Invalid y^x". Still, the HP 32 SII is a rather good calculator.

Since Garmin now have requested my Vista back, to investigate the "Anders bug", maybe I should click to the year 2400 and see if it's correctly handled? Not that I think I'll live to be bothered by a potential fault there...

Anders

[This message was edited by Anders on August 19, 2002 at 11:02 PM.]

2.3k · August 19, 2002

There were several Commodore calculators for sale here, but that was before the Commodore 20 and 64 computers. We are talking 1975-1980 now. I still have a few of them in a box in my basement.

My Hewlett-Packard 48 SX says 0^0 is 1, but the HP 32 SII says that's an "Invalid y^x". Still, the HP 32 SII is a rather good calculator.

Since Garmin now have requested my Vista back, to investigate the "Anders bug", maybe I should click to the year 2400 and see if it's correctly handled? Not that I think I'll live to be bothered by a potential fault there...

Anders

[This message was edited by Anders on August 19, 2002 at 11:02 PM.]

August 20, 2002

http://mathforum.org/library/drmath/view/55764.html

that result seems correct according to this URL :-)

August 20, 2002

OK, perhaps the calculator isn't incorrect with respect to 0^0. It's apparently a still-debated mathematical topic, see http://mathforum.org/dr.math/faq/faq.0.to.0.power.html

My HP48GX agrees with your HP48SX and the Microsoft Windows Calculator (v4.0) in giving 1. However, MS Excel returns an error. I don't have Maple or Mathematica handy to see how they weigh in.

In any case, there is an argument to be made for either answer; 1 or undefined/error.

df

August 20, 2002

At least that's the way I learned it those many years ago in math class. My HP48SX agrees with me.

quote:
Originally posted by DarrenF:
I don't have Maple or Mathematica handy to see how they weigh in.

Mathcad, which uses Maple for calculations, also agrees with me that the result is 1. But Mathematica calls it indeterminate.

Until forced to believe otherwise I'll continue calculating anything to the power of zero as one.

I'm not lost!

I just don't know where I am.

August 20, 2002

At least that's the way I learned it those many years ago in math class. My HP48SX agrees with me.

quote:
Originally posted by DarrenF:
I don't have Maple or Mathematica handy to see how they weigh in.

Mathcad, which uses Maple for calculations, also agrees with me that the result is 1. But Mathematica calls it indeterminate.

Until forced to believe otherwise I'll continue calculating anything to the power of zero as one.

I'm not lost!

I just don't know where I am.

2.3k · August 22, 2002

The calculator in release 2.29, for the Vista, has improved. Error messages for ln(x), where x<=0, exponents 10 and 20 correctly shown and so on. 0^0 still results in an ERROR!, but it seems we can accept that, can't we?

Anders

2.3k · August 22, 2002

The calculator in release 2.29, for the Vista, has improved. Error messages for ln(x), where x<=0, exponents 10 and 20 correctly shown and so on. 0^0 still results in an ERROR!, but it seems we can accept that, can't we?

Anders

2.9k · August 25, 2002

0 ^ 0 = Ma ERROR on my Casio Scientific Calc.

if 0^0=1 then god created the universe out of nothing.

------------------------

My home page about GPS units and information

August 25, 2002

quote:
Mathcad, which uses Maple for calculations, also agrees with me that the result is 1.

Odd... I'm using Maple V Release 4 right now and it says "Error, 0^0 is undefined"

df

2.3k · August 27, 2002

Since the 0^0 debate seems to have no real (pun!) solution, try this instead:

Enter 4, then square that five times. Seems reasonable, right? Square it once more, and you get what you could expect.

But then do the same, starting with 5, 6 and 7.

Oops!

Anders

2.3k · August 27, 2002

Since the 0^0 debate seems to have no real (pun!) solution, try this instead:

Enter 4, then square that five times. Seems reasonable, right? Square it once more, and you get what you could expect.

But then do the same, starting with 5, 6 and 7.

Oops!

Anders

August 27, 2002

quote:
Originally posted by Anders:
Since the 0^0 debate seems to have no real (pun!) solution, try this instead:

Enter 4, then square that five times. Seems reasonable, right? Square it once more, and you get what you could expect.

But then do the same, starting with 5, 6 and 7.

Oops!

Anders

All of those seem to work perfectly well in the Legend 2.40c calculator.

df

August 27, 2002

quote:
Originally posted by Anders:
Since the 0^0 debate seems to have no real (pun!) solution, try this instead:

Enter 4, then square that five times. Seems reasonable, right? Square it once more, and you get what you could expect.

But then do the same, starting with 5, 6 and 7.

Oops!

Anders

All of those seem to work perfectly well in the Legend 2.40c calculator.

df

August 27, 2002

(again, using the calc in the Legend 2.40c firmware)

143^143 = ERROR!

144^144 = 0

And,

138^144 = ERROR!

139^144 = 0

df

August 27, 2002

quote:
Originally posted by DarrenF:
(again, using the calc in the Legend 2.40c firmware)

143^143 = ERROR!

144^144 = 0

And,

138^144 = ERROR!

139^144 = 0

df

Wow. That zero (the result of the above 139^144) is NOT the same as any other "0".

I did 144^144, got "0", hit "ln" and got -722.627286

So I thought, "that's wacky, wonder if what if I try to invert 'that zero'"...

I hit "C", did 144^144, got "0", hit "1/X"... and the GPS shut down Repeated... same result.

Looking at the Disgnostics (Trip Computer, zoom-out, zoom-in, zoom-out), it lists the shutdown as an "RTL Trap"

r0: 0000f504 r8: 00000000

r1: ffffffff r9: 00000000

r2: 040355d0 r10:00000000

r3: 64800000 r11:00000000

r4: b1321ccd r12:00000004

r5: 040077f0 r13:04006930

r6: 00000000 r14:0201d56d

r7: 04007774 r15:0201d569

errn: 80000201

Appears to be a CPU register dump and error number (?)

df

August 27, 2002

quote:
Originally posted by DarrenF:
(again, using the calc in the Legend 2.40c firmware)

143^143 = ERROR!

144^144 = 0

And,

138^144 = ERROR!

139^144 = 0

df

Wow. That zero (the result of the above 139^144) is NOT the same as any other "0".

I did 144^144, got "0", hit "ln" and got -722.627286

So I thought, "that's wacky, wonder if what if I try to invert 'that zero'"...

I hit "C", did 144^144, got "0", hit "1/X"... and the GPS shut down Repeated... same result.

Looking at the Disgnostics (Trip Computer, zoom-out, zoom-in, zoom-out), it lists the shutdown as an "RTL Trap"

r0: 0000f504 r8: 00000000

r1: ffffffff r9: 00000000

r2: 040355d0 r10:00000000

r3: 64800000 r11:00000000

r4: b1321ccd r12:00000004

r5: 040077f0 r13:04006930

r6: 00000000 r14:0201d56d

r7: 04007774 r15:0201d569

errn: 80000201

Appears to be a CPU register dump and error number (?)

df

2.3k · August 27, 2002

Well, this is what I get on my Vista:

Enter 4, then square five times. Result 1.84467E+19.

Square once more. Result: ERROR!

Just what one could expect.

Enter 5, then square that four times. Result: 152587890625. So far so good. But square once more. Now it should be 2.32831E+22, but it shows 9.005009E-308, which most certainly isn't correct.

Starting with six, I get 2.821110E+12 when squared four times, but 1.398044E-76 after yet another square.

Is this on the Vista only, or is it one of the infamous "Anders Bugs" that's come back, under a different disguise?

Anders

August 27, 2002

quote:
Enter 5, then square that four times. Result: 152587890625. So far so good. But square once more. Now it should be 2.32831E+22, but it shows 9.005009E-308, which most certainly isn't correct.

Using Legend 2.40c I get 2.328306E+22, one more square gives ERROR!, as expected.

quote:
Starting with six, I get 2.821110E+12 when squared four times, but 1.398044E-76 after yet another square.

I get 7.958661E+24 for that operation.

df

2.3k · August 27, 2002

I just read your post, Darren, and found that 144^144 = 1/x turns off my Vista too.

It also generates a RTL trap diagnsotic, where the error number in the dump is the same. Register contents isn't, of course, since that's likely to be dependent on other things you do.

I had two more RTL traps (seems to be something new, haven't seen them before), probably as a result of the unit turning off when I played with the new Jump Master accessory. When I pressed "Goto DIP", it happened twice that it shut down.

Anders

August 27, 2002

I can confirm that the ln(x) for non-positive x now correctly returns an error, as does 1/x for x=0. Scientific notation results with an exponent of 10 or 20 are also now correctly displayed (woohoo for fixed bugs).

The "144^144=0 followed by 1/X" shutdown behavior is still present.

This version also has a very strange tendancy (that was not present before) to give results in scientific notation. Almost any calculation that returns a non-intiger results is given in scientific notation(!). Any other Legend v2.41 upgraders notice this? Does the new Vista firmware's calculator do the same?

df

August 27, 2002

quote:
This version also has a very strange tendancy (that was not present before) to give results in scientific notation. Almost any calculation that returns a non-intiger results is given in scientific notation(!). Any other Legend v2.41 upgraders notice this?

Yes, sceintific notation seems to be used for many results, but not all. Some decimal results are still decimals ( 100 and 1/X gives 0.01 ) but the same result through different means yields a scientific notation ( 1 - 0.99 = 1.000000E-02 ). It may be that non-integer arithmetic results always yeilds scientific notation.

It also still has the odd result 5 ^ 32 =9.005009E-308. And 1/X of that result is 0.03125, which is exactly 1/32. Odd.

(EDIT) now when I try 1/X of E-308 result it is just an error, must have been a fluke the first time - calculated on the 32 from 5^32 (/EDIT)

I'm not lost!

I just don't know where I am.

2.3k · August 27, 2002

This scientific notation preference is already in the Vista. When Darren gave his example of 144^144=, then ln(x) of the result, I got the same numeric result on my Vista, but in scientific notation.

Anders

2.3k · August 28, 2002

Interesting to see that 18^18 = 5^32.

And also that 19^19 = 6^32.

All this according to the Vista calculator.

The first line gives the answer 9.005005E-308, the latter 1.398044E-76.

No, none of the two are actually equal, and most certainly not as small as these values suggest.

Anders

2.3k · September 24, 2002

Whoever is responsible for that calculator accessory at Garmin is doing his homework (probably math!).

The 2.29d beta release to the Vista has cured just about every known problem with the calculator.

One can still enter many decimal periods in succession, but I didn't know about that either, until I happened to do an error the other day.

If they now could get an inverse key, for the trigonometry and the ln key, into that scientific calculator, it would be really useful.

Or even better, if they could get rid of some other bugs and fix som more GPS related features.

Anders

Testing the calculator on my eTrex Legend with firmware v2.40c

Recommended Posts

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Link to comment

Join the conversation