Speed differences between MC and Rev and the origin of the English language

[Wilhelm Sanke]

Home again and back to work as I announced 11 days ago. There is a 
350-pages doctoral dissertation on my desk - a study about internet use 
in schools - which I need to assess and grade during the next two weeks, 
but I resume paying some attention to my imagedata stuff and MC.

My two vacations this summer were spent in parts of Germany where 
language minorities still fight against assimilation. Last week I 
visited the Sorbian area near Poland where a small number of people 
still speak their old Slavonic language which is also still used in 
schools. In July I spent a quiet time in the northernmost county of 
Germany, called "land Angeln" since 2000 years - South of the Danish 
border - from where the English language and especially the name for the 
"English" language originated.
I had taken my laptop with me to the land "Angeln" to work on my 
"imagedate toolkit", but in the first night the harddisk crashed and I 
had two weeks time to attend to other matters and study the local 
culture and what had remained from old times; additionaly I consulted my 
memory and later Wikipedia and my old textbooks. As you all on this list 
are more or less familiar with the English language it will not hurt to 
share my collected insights about its origin:

The short story of the development - and about 99% of this is true - 
runs like this (and the Scots are also instrumental here, if only in a 
somewhat negative way):

The Roman emperor Hadrian had built his Hadrian's wall across England 
from Newcastle to Carlisle because he did not like bagpipe music and 
kilts. When the Roman empire broke down in the 4th century, the 
revolutionaries from Edinburgh again annoyed the Celtic population South 
of the Hadrian's wall. Then the British Celts asked two former Roman 
mercenaries, namely Hengist and Horsa, to help them against the Scots. 
Hengist and Horsa happened to belong to the Anglian tribe in the land 
"Angeln" and asked their relatives and more tribesmen over to Britain. 
The Saxons - South of the land "Angeln" - and the Jutes to the North 
joined them and established their kingdoms (and languages) in the new 
country. The last Celtic king to fight against this invasion was the 
legendary King Arthur from Tintagel in Cornwall.

Later other waves of immigrants came over to England from the same 
places where the Anglians, the Saxons, and the Jutes originally lived. 
 From the eight century on the Vikings made frequents inroads  and then 
settled in England (the famous "Hägar" was one of them). In about 10 
miles distance from where I spent my July vacation is "Haithabu", the 
former trade center of the Vikings in the land "Angeln" (a place similar 
to Sutton Hoo in England, where big Viking ships have also been found). 
In the 11th century after Hastings the Normans, which were originally 
Vikings, too, came over who had made a detour through the Normandy in 
France, got civilized there and had learned some French and French 
cuisine in the meantime. Then even later more Northern Germanic tribes 
invaded, especially Jutes (Danes) who established the "Danelag" in 
England. Knud the Great was at the same time King of England and 
Angeln/Denmark/Norway. All these Germanic tribes spoke closely related 
Germanic dialects very similar to the language documented in "Beowulf", 
the oldest extant example of poetic "Old-English" language. 
Interestingly, the venue of the saga told in Beowulf, this center piece 
of Old-English literature, is the Southern area of Scandinavia.  

By the time of Chaucer the Anglian/English language had evolved as the 
widely spoken and officially used means of communication in the new 
"England". It is an astonishing fact of language development that the 
term "Anglian/English" , derived from the small land-Angeln region of 
continental Europe, finally prevailed as the name for one the most 
important languages and the "lingua franca" for international 
communication of today. At least the Saxon part of this language 
development is honored when we sometimes also speak of Anglo-Saxons and 
Anglo-Saxon languages.--

======================

Now to the speed differences:

On Sat, 09 Sep 2006, Richard Gaskin <ambassador at fourthworld.com> had 
written (Subject: Re: [ANN]: Imagedata Toolkit (beta) released):

> Wilhelm Sanke wrote:
>
> >> (Remark for Metacard list members only:
> >> For optimal performance use the Metacard IDE. When run in the 
> Revolution
> >> IDE some filters need up to 45% more processing time than in MC. This
> >> also holds when you create a standalone.)
>
> How could that be possible when they both use the same engine?


The Rev IDE is a slowly evolving environment with about 20 times more 
code than in the MC IDE. Given the additional intricacy and 
interrelatedness of this code it is quite natural that speed and other 
problems occur. Some of my earlier stacks - the discussion about this 
and examples are still to be found on my website (I will remove them 
soon) - would not even start in the Rev IDE or crash or could not be 
inspected by the application browser or property inspectors because 
these tools were unable to handle larger number of objects in a stack. 
The Rev IDE indeed has considerably improved over time, but IMHO has 
still a while to go to become a reliable and more user-friendly application.

As to standalones, apparently something is being added to them in Rev 
what is happily missing in Metacard.

Richard,
six months ago - on March 1 - I had sent an offlist post to you 
concerning the same topic. I will quote from this post for the benefit 
of the other list members:

> > Did you ever determine the cause of these performance differences?
>
> Hello Richard,
>
> I repeated the comparison of  MC and Rev IDEs and standalones with the 
> latest version of 2.7. GM 1.
>
> As testscript I used the duplicating of imagedata colors of a 640X480 
> image. The image returns to its original appearance after 8 runs of 
> the script. Each run needs a slightly different time, because the 
> number of corrections for each color value necessary when the 
> duplicated color value is greater than 255 - their is an if-statement 
> that takes care of that - is different for each run.
>
> The speed varies between 2355 and 2779 milliseconds with MC and 2849 
> and 3335 in Rev, the average speed difference between MC and Rev is 
> about 600 milliseconds (2 GHz machine). Running the test in the IDEs 
> or as standalones makes no apparent difference.


For the latest IDE versions with engine 2.7.3 and with improved scripts 
these values are now like this:

An almost constant speed of slightly below 1700 milliseconds for MC and 
an equally nearly constant speed of  2350 milliseconds in Rev, but with 
a peak of 2400 milliseconds. Building a standalone in both IDEs indeed 
does not change these results.

You can test this using sample images "Portrait of the Artist" and "Red 
Square Moscow" from my Imagedata Toolkit (to be dowloaded from my 
website <http://www.sanke.org/MetaMedia>) and using filter "Duplicate 
Colors 1". "Red Square Moscow" provides better results in terms of 
interesting colors.

Continued quote of my post from March 1:

> I saw to it - as I explained in my last post - that no inclusions or 
> any libraries were added to the standalones.  However the resulting
> Rev  standalone is somewhat bigger than the MC one, about 150 KByte.
>
> As substack "revsaveasstandalone" of stack "revStandaloneSettings" is 
> password-protected I cannot find out what happens here; they should 
> definitely remove this password protection (the necessity for this - 
> as far as I remember - was to enforce an additional logo that appeared 
> when a standalone was closed indicating with version of Revolution was 
> used -"Studio", "Enterprise", "Educational" etc.).
>
> If the password protection would be removed RunRev would gain the 
> opportunity to get help from interested member of the community to 
> improve the Rev standalone builder.-
>
> By the way, I also used Chipp Walters "altclean"  plugin with the Rev 
> stack to remove unnessary components of that may have been inserted 
> when working in the Rev IDE, but this did not change the test results.
>
> (snip)
>
> Wilhelm Sanke


Best regards,

Wilhelm Sanke
<http://www.sanke.org/MetaMedia>