In the last years, several scholars - an even some Rroma - have advocated a standardisation of Rromanes - to coalesce the language into a single variant. In the view of its proponents, this standardisation is a mean by which a true Rroma trans-national identity can be forged.
So what can be done and what should be done. First, one needs to keep in mind that Rroma are by nature dual-language speakers. We will describe this more extensively in the section about the Nordic group to show how this influences the creation of groups-dialects. All Rroma who speak Rromanes also all speak the language of the country they live in and in many cases even some additional languages. As a result, a strong influence and interference which is totally natural and local linguistic peculiarities "creep" into Rromanes - as we have seen from the Rroma history.
Some attempts to create a unified standardised Rromanes have almost collapsed into ridicule. The aim of creating new lexems by miss-placed "neologisms" can sometimes bring amusing side effects. In most Rromanes dialects, the suffix -lin is connected with fruit trees such as phabelin [apple tree] from phabaj [apple]; ambrolin [pear tree] from ambrol [pear] and so on.
We have come across an attempt to name a kindergarten a xurdelin, from the Rromanes xurde with a dual meaning: First and for most Rroma this means small change, i.e. money, second, in some dialects it is used for small children. Whether the children tree or the money tree is more understandable to Rroma as a kindergarten remains to be seen - although the latter meaning as a money tree may be seen as a possible way to entice Rroma to send their children to such a place...
Adding to this plight, some literal translations from other languages, although perfectly grammatically correct can gain a totally different meaning. In a book for young Rroma children and distributed in various countries with the help of some international organisations, one can find the following sentence about a small dog called Rukun "Rukun so tu andan and-i škola? - Mo kokalo." meant to mean "Rukun, what did you bring to school? My bone". Unfortunately for the writer most Rroma (from our personal experience) understand not that the dog brought a bone to eat to school but that he brought his own bones, his own skeleton.
The last if not the worst is the attempt in this particular case to create new letters used only in some grammatical forms.
The French ç of this new alphabet is used exclusively in the forms instrumental cases such as in Rromeçar for Rromesa(r) [with a Rrom]. When asked about the logic behind this use of ç the author justified himself by saying, as we have seen above that there are several dialectical variations such as Rromesa, Rromeha, Rromeja and thus, this could be used for all of them. Unfortunately, in the cases of the verb to be, he uses the simple si although one sees the same variety in this verb as the one hinted earlier.
But there are many variants such as hi that exist in many dialects. In the case of this reflection between h and s, one should be consequent and thus use the ç in all cases where it occurs such as in sar / har, savo / havo, and even so / ho. The same is true of the inclusion of the q in the dative case: Rromesqe [to the Rrom]. The logic is the same as variations such as Rromesče are found in some dialects or regions. Kerav [to do] which shows same reflections i.e. čerav is however always written with a k in this "new alphabet".
Anyhow, alphabet will remain a hurdle towards such attempts at a standardisation. In Many countries, the only known alphabet is the local one, such as the Cyrillic one in Russia, Ukraine, Byelorussia, Bulgaria, Macedonia and Serbia, complicating these matters.
In terms of spelling, we cannot but cite the difficulties encountered by Dieter Hallwachs about which writing to use in a newly created orthography of the Burgenland Carpathian Rromanes. The Rroma living in Austria vehemently opposed the inclusion of such letters such as š or ž preferring their German spelling of sch. As in the case of Romantch in Switzerland, one can ask why a standardisation on one model is necessary. Romantch was reintroduced in Switzerland in 3 main variants, each corresponding to a geographical entity - basically a valley. Schools, newspapers use these three forms which correspond to the language spoken at home. For Rromanes, one could and should advocate the natural affinities of the dialects of the different groups to slowly forge "metadialects" which have a broad base. Then, with time, one may hope that a true "standardised" Rromanes emerges.
Upon imposing one or the other dialect (or, as we have seen, an artificial one) as the basis of the standardised language, one runs the problem that it will neither be understood nor accepted. Beside, in the previous section, we have asserted that no dialect is "better" than another one. This means that one cannot find a conventional form of Rromanes upon which such a standardisation could be built.
Needless to say that schooling is also almost useless since Rromanes spoken at home will always prevail over artificial implants.
The right and only way to proceed is thus of course to restrict oneself to the broad dialect classes - which we call metadialects. This could create a "common" base inside a given dialect group and eventually lead to a unified language. Once again, we stress that this process only has a chance of success if the evolution comes from the community inside a dialect group and none at all if it is imposed from above