A new variable scope

[forskare]

Re: A new variable scope

Hello Ira,

From what was posted, I understood everything you said. I don't think Gabe argued that it would work in a multi-user environment. However, it appeared to me that some were poo pooing his entire theory when it appears to me it would work in a single user environment. In fact, some really didn't seem to understand what he said. That is why I wrote my first response.

My comment, "the jury is still out" referred to the single user environment. I should have point clear.

Thank you for taking the time to respond.

kenn

[G Gabriel]

Re: A new variable scope

Ira:
I am glad you had a revelation and found your seven keys because in your ealier posts you said:

you could also store data in windows environment variables, but this memory space goes away when you shutdown windows, and also very often is not configured for sufficient space to hold much data in many (maybe most) computers.

It would be nice if Alpha had provided an "Atomic" locking mechanism for code, but it does not, and that is why it is so difficult.

and denoucing processing in RAM drive:

Actually, I have. And surprisingly, Alpha 5 ran slower than with the hard drive. Explaining this is difficult, when it seems intuitively obvious that it should be the other way. I had tried several Ram disk drivers with similar results

Unfortunately, Alpha 4, eide drives (and older OS's?) does not compare to DMA drives (now standard, although it wasn't in the past), Gigabytes of memory, extremely fast hard drives with caches etc. So it needs a revisit by many under various environments to validate my results. I'd love for it to turn out in Ram disks favor, but based upon my tests, wishing will not make it so.

though the issue is simply updating a single variable.

Ira, I don't wish to keep this up because from so many encounters in the past I have come to the conclusion that you and all members of the GOP will never ever acknowledge the fact that maybe just maybe someone thought of something you did not think of first. It is quite clear that you are very knowledgeable, I give you that. At least I have the decency to admit that and further admit that I am nowhere near as knowledgeable about programming as you are. Why? Because that's not what I have been doing for the last few decades. But what you are not able to see or admit, at least to yourself, is that knowledge in and by itself is not sufficient, and that innovation is not directly propertionate to knowledge or merely and solely based on the quantity of knowledge.

Now as we all know the laws of physics always apply. To every action there is a reaction. When each and every time you go on trying to shoot down my plane before even understanding fully what I am talking about, guess what? I am not going to fly too many planes. Hope that makes you happy. And also as a matter of reaction to your and others' actions, I have recently posted several zip file examples without showing the code. I am waiting for you or any member of the GOP to match it. You haven't! It's been a while now, so are you going to put all this knowledge to good use and match these examples? Or are we going to hear way too many information that does not answer any questions?

In case you wonder what is the GOP, it's the "Gabriel Opposition Party".

[MikeC]

Re: A new variable scope

G,

I have long touted you for being highly knowledgable and an innovative thinker...and have even supported you on occasion but enough is enough. Whatever respect I may have had for you is gone. For you to "tease" people who have no direct part in your "reaction", by leaving out the method in which you have apparently (I say apparently as without seeing what was done does not show it was done per request) solved the problem is unwarranted and just plain mean spirited.

Just stop. I, for one, can do without the knowledge that comes with such a high price as you demand.

[G Gabriel]

Re: A new variable scope

It's not mean spirited nor meant to be. I have always offered many examples in the past freely and without any expectations. But this none stop assail is getting to me. I am tired of having to keep responding to meaningless debates and was only hoping to put an end to this once and for all.

[MikeC]

Re: A new variable scope

Yes G, but something changed--don't you remember that you deleted ALL your samples a while ago? All the images, zip files, etc which leaves all your past helpfulness nonexistant really now doesn't it? And I expect the same again in the near future...a regular pattern is emerging. Too bad really. I really was hoping your last tirade was not going to be repeated but see no indication of that by your recent actions.

I'm sure many will agree that "a mind is a terrible thing to waste".

I only responded again to make sure any newcomers realize just what you have done...and what you continue to do... so that their expectations of you are of something that you are able to accommodate.

[Al Buchholz]

Re: A new variable scope

It's not mean spirited nor meant to be. I have always offered many examples in the past freely and without any expectations. But this none stop assail is getting to me. I am tired of having to keep responding to meaningless debates and was only hoping to put an end to this once and for all.

You initialize that with the first step that you take inward.

[G Gabriel]

Re: A new variable scope

Yes G, but something changed--don't you remember that you deleted ALL your samples a while ago? All the images, zip files, etc which leaves all your past helpfulness nonexistant really now doesn't it?

No not really. Whenever the occasion rose, I redid them. Plus, these are things from the past; there are many newer ways to do things that I have adopted since. And further, my tirade was mainly directed towards alpha. Mike, if you are going to come so strong condemning me and "exposing me" to the whole world, at least try to get the facts first.

Why was I mad at alpha?
Unlike many, I don't just go on a rampage reflexively. But when I get fed up, I do. There were several occasions where alpha deleted my posts where I felt it was inappropriate of them to do so. To give you a simple example, in one thread two Croatian guys were going back and forth talking to each other in Croatian. I said this is an English-Speaking board, please put your replies in English. They continued to chat in Croatian. I wrote: enough of this already! And sent an email to alpha to remove that thread. We don't know what these guys are saying, do we? Instead, Alpha deleted my post and kept theirs! Well, that didn't bother me all that much, but then in few other threads, they did the same thing deleting my responses and when you read the thread after the deletion you come to the conclusion that I am an idiot.

So, in my response to alpha I basically said, OK, If you are going to selectively delete my replies, at least do the honorable thing and delete the whole thing or I will do it for you.

I later received an apology from alpha in a private message!

[G Gabriel]

Re: A new variable scope

You initialize that with the first step that you take inward.

OK Al. Of course and as this thread and others show, I am the one to blame. This thread could have been over many moons ago!

[forskare]

Re: A new variable scope

I have particularly enjoyed listening to well respected developers argue points back and forth. To quote Mike C in a comment to Gabe, "Always enjoy your somewhat sideways twist on things...many times being just so simplistic that others skim over without even seeing them." That is very true in this instance.

None the less, I thought it was a great discussion, at least until the comment, 'Dead embrace' was made. I wasn't sure of the meaning but I took it to mean the author was making a point that Gabe was embracing a point of view that would go nowhere. That was not in good taste and pushed Gabe's button.

I also think some comments were made with a bit of attitude towards Gabe. I'm not saying the attitude is not justified. What I am saying is this: We ought to be careful that we respond to the issue and not the person. Attacking the person only breeds ill feelings as this and past threads have proved. Some are still harboring ill feelings which, when not dealt with in a positive manner, will only fester and at some point, resurface in an even uglier manner as it did in this thread.

In summary, I see positives and negatives on both sides, not just with the subject but in how both sides responded. Sometimes, because of our personalities, we say things in a tone that others see as offensive but we have no clue.

It would be most helpful to all who participate in this forum, if those involved in this thread, myself included, would take the time to carefully review this thread and genuinely ask themselves, "Could it be that I said something that offended another, even though I didn't think I did or I believed I was justified"?. For me, if anyone believes I have said something to offend, please send me a private email and let me know what it is and why. It WILL be dealt with in a professional manner.

Thanks,

kenn

[G Gabriel]

Re: A new variable scope

Ken:
I believe deadly embrace in computer jargon refers to a situation where two scripts hogs the table at the same time.

Here is a better definition than my chicken leg definition:

An example of a deadlock occurs frequently in database products. Client applications using the database may require exclusive access to a given table, and in order to gain exclusive access they ask for a lock. If two client applications both attempt to lock the same table at the same time, neither will receive it, there is no general way to decide who to give the lock to. In this case both clients will wait for the lock forever.

A slightly different definition describes when two tables are locked sumiltaneously.

It's like when you walk down the street and come face to face with someone and you turn one way to avoid him while he turns the same way, then you turn the other way and he does the same thing and it goes on indefinetly.

The interesting part is deadly embrace if it occurs it will occur with physical tables, NOT when updating a variable in memory (although concievably it can), less argument for using tables. Will it occur with attempting to save a script? maybe, but I doubt it. If right before saving the script you set a global flag to off, that would prevent the other user attempting to save at the same exact nanosecond from saving. What if the two users set the flag to off at the same nanosecond? neither will be able to save but then you could re-attmept at intervals and if you make the intervals different from user to user even by split seconds, that ought to prevent this locking.

Ready for more aesoteric terms? how about "conversion lock"

[csda1]

Re: A new variable scope

Hi Kenn,

II thought it was a great discussion, at least until the comment, 'Dead embrace' was made.

I believe John was expressing his delight in learning a new computer term of "Deadly Embrace"

Deadly embrace is locking issues taken to the next level (locking is the process of handing off 1 resource without regard for other resources). It is where one process (a session or thread in Alpha 5) is using 1 or more resources (e.g. a table record in change mode), and requires a resource (such as a child record) that another process has control of. In the mean time, that 2nd process requires a resource that the 1st process already has. They both wait forever for a resource that is never coming.

And as you add more processes to the mix, each could be waiting for resources from others and all end up waiting. Nothing gets done

This is similar to sending packets on the network. If everyone keeps trying to send, no data gets through. But if each gives up their use of the network for a random time, they all have a chance of getting their data through on the network.

The economic genius John Nash's theories also have key applicability to this.

Under controlled circumstances, deadly embrace can be handled properly, but never solved generically.

[forskare]

Re: A new variable scope

Well, what do ya know about that? I'd not heard that term before and not doubt there are at least a few others. So, I owe John Zan apology.

"John, I apologize for not getting clarification of what you meant by Deadly Embrace. Used in the context of this thread, it is very appropriate."

kenn

[DaveM]

Re: A new variable scope

I think this thread could be used to pump Alpha toward a newer way for them to alocate some memory from the server environment. It would have to be a seperate memory entity than the local environment as I see it.

I just hope some of the Alpha people see this and further hope it can be done. It would be a novel way to get the results G describes.

In Short, I LIKE it!!!

what would be wrong with loading a script into a memory space? I believe I have heard Alpha reads a script repeatedly in some situations. Yes it is fast, but that would be faster.

Dave

[csda1]

Re: A new variable scope

Hi Dave,

I think this thread could be used to pump Alpha toward a newer way for them to alocate some memory from the server environment. It would have to be a seperate memory entity than the local environment as I see it.

I just hope some of the Alpha people see this and further hope it can be done.

Functions to do this are already there (although not quite as simple as declaring a variable and using it), although with the downside of modifying a data dictionary, IMHO, a bad idea in general. And if you simply distribute updated data dictionaries, those data dictionary values are toast.

So if you'd like to dynamically mess with your data dictionaries, the functions include

A5.WRITE_SETTING( Setting_Name as C, Value as C )
A5.READ_SETTING( Setting_Name as C )

and it's big brother versions

A5_LOAD_SETTINGS( Setting_Name as C, Dot_Variable as P [, Dictionary_Name as C ] )
A5_SAVE_SETTINGS( Setting_Name as C, Dot_Variable as P [, Dictionary as C ] )
A5_DELETE_SETTINGS( Setting_Name as C [, Dictionary as C ])

If you'd like to mess up your Windows registry, we also have these functions (although at least these are only available to a single user)

REGISTRY.LOAD_SETTINGS( Registry_Key as C, Dot_Variable as P )
REGISTRY.SAVE_SETTINGS( Registry_Key as C, Dot_Variable as C )
REGISTRY.DROP_SETTINGS( Registry_Key as C )

But if you really want to just waste more time, and slow things down, then use

SCRIPT_LOAD( Script_Name as C [, Password as C ] )
SCRIPT_SAVE(C script_name,C script_code[,C Password])

which have a 48 byte automatic overhead in size for even a zero length of code, and have a much higher overhead in the read and write process to a data dictionary (when you write the script, Alpha attempts a compile of the script code which accounts for the long write times, but loads are slower too).

And all of the above are not locked during the change process (you can't keep them in change mode while reading the current value as in a table), which means they absolutely will all eventually fail in any normal multi-user (or multi-process) environment.

Or you could do the identical type operations with a special data table, and be a whole lot faster, multi-user and safer from corruption. I know what I'd choose.

what would be wrong with loading a script into a memory space? I believe I have heard Alpha reads a script repeatedly in some situations. Yes it is fast, but that would be faster.

Alpha 5 already caches script code execution in memory. But if you use scripts like data, you potentially invalidate the mechanism that does the caching, as data caching is different from "read-only" type caching (but it depends on Alpha Five's caching algorithm). Anyway you look at it though, you cause more cache flushing, slowing all of Alpha's script execution.

[G Gabriel]

Re: A new variable scope

And all of the above are not locked during the change process (you can't keep them in change mode while reading the current value as in a table),

Ira:
Let me repeat in case you missed it:
You do not need to load the script in order to save the new one.