If Arabs were - supposedly some day - have taken their numerals from Indian Devanagari script or even another one by the end of 796 AD (the death-date of Alfazari), numerals' forms, as appeared in the past first two manuscripts, would not have been that different from the nearest of the Indian scripts I have exhibited within only 300 years of the time of Mohammad Ben-Ebraheem Alfazari, and less than 400 years from the reign of Umayyad Caliph, Abdolmalek Ben-Marwan, unless that (some day) is deep in the antiquity thousands of years before the date of a manuscript like: "Mokhtasar Alhendi Fi Alhesab" [Conclusion of Alhendi on Arethmatic], which had been written before 1100 AD.
Yet, there are still other proofs to the Arabic-originality of the numerals which I see its rise to may have been from the Arabic names' abbreviations of the numerals. Following is an analysis for the outline transformation that has come over the Arabic prototype numerals (extrapolated), branching them into Arabic-Eastern and Arabic-Western (Ghebari) forms:
As for the Arabic digits seven and eight, it is most probably that their ancient forms have been written in an advanced phase of the prototype Arabic numerals to be close from that of digit seven as in Tumbuktu's manuscript , something like these following seven and eight consecutively: . Then, to avoid confusion between these two numerals' forms in the west of the Arabic World (in Ghebari numerals), the tail of the digit eight has been turned to the right as such: , so that it ends up to the eight digit's form as shown in Tumbuktu's manuscript: ; whereas the Arabic-Easterners have oriented each digit's face in an opposite direction to the other so that we get the two forms: ٨ , ٧ of them in the Arabic-East.
At this point, we need to know the reason for the Arabic numerals' transformation into Arabic-Eastern from its ancient (Ghebari) form.
The expanse of the Islamic-Arabic World's area horizontally from west-of-China eastwards to the Atlantic ocean westwards through the northern half of the African continent - in a time when means of transport remained crawling type-wise and speed-wise on one hand, and the ethnic multiplicity of the Islamic-Arabic World (especially in its east) on the other hand - necessitated the varying and diversifying of brainchildren and arts so that a scripter or a copyist in the Arabic-East draws differently than his counterpart in the Arabic-West.
This variation has not, of course, been limited to drawing the Arabic numerals, but clearly appeared in writing the Arabic alphabet, so the same Arabic alphabet for the same language, even - for instance - for the same Qor-ann in Iraq is written in some style/type (Kufi calligraphy), whereas in the Arabic-West and Andalusia is varyingly written (Andalusi calligraphy). These are not but just two of several instances of Arabic calligraphies. Even the architecture, no wonder it varied from place to another, as we can see a mosque built in Persia at a particular style, while you see that that is built in Tunisia or Andalusia, or even in Damascus and San'a, including the same structure but most often varies in the appearance and look.
The transformation in the form of the Arabic-Eastern numerals is a bit clearer and wider than it in the Arabic Ghebari in the Arabic-West, from which Arabic numerals transmitted to Andalusia and Sicily, then to the rest of Europe, for two reasons:
The first is the high population and ethnicity dense in the Arab-Islamic-East than it in the West which is confined by the African great Sahara (From Libya and Chad to the Atlantic Ocean). The second is the high human activity and dynamism in the Arab-Islamic-East than it in the West. The political hostility - for some time - between the Islamic state in the East and its counterpart in the West might have represented an additional reason for the separation between the two forms of the Arabic numerals in the two parts.
It is noticed that transformation in languages or scripts runs faster and wider where the human activity is greater in the variant sides of life. We see plenty of changes have come over the French language in France, while the French of Quebec's inhabitants in Canada remained more conserving to its past, clearly differentiating between them. One other example has happened with the Portuguese in its native land, Portugal, than that being current in Brazil, and is referred to as the Old Portuguese. The separation - in at least these two examples - has particularly run under the same conditions which are: the far distance between the two parts; and the weak communication and its means in that ancient time.